Skip to content

Choosing a dataset

flowchart TD
    A[What do you need?] --> B{Fire happening now?}
    B -- Yes --> C[FIRMS, GOES-R, NOAA HMS, FEDS]
    B -- No --> D{Past fire extent?}
    D -- Yes --> E{Where?}
    E -- US --> F[MTBS, WFIGS, FRAP, Alaska Fire History]
    E -- Canada --> G[NBAC, CNFDB]
    E -- Global --> H[MCD64A1, VNP64A1, Fire_cci, GlobFire]
    D -- No --> I{Fire drivers?}
    I -- Weather --> J[gridMET, RAWS, HRRR, Drought Monitor]
    I -- Fuels --> K[LANDFIRE, Globe-LFMC, TreeMap]
    I -- People --> L[GlobPOP, CDC SVI, WUI maps]
Question Start with Theme
Where is a fire burning right now? FIRMS, GOES-R, NOAA HMS Active fire
How big was a past fire and where did it burn? MTBS, WFIGS, FRAP, NBAC, MCD64A1, VNP64A1 Perimeters
How many fires, what caused them? FPA-FOD, EFFIS Occurrence
What will burn and how fast? LANDFIRE, gridMET, RAWS, Globe-LFMC, US Drought Monitor Fuels, Weather
How bad is the smoke? GFED, GFAS, AirNow Smoke
Which homes were lost? CAL FIRE DINS Impacts
Where was prescribed fire used? SE FireMap, FACTS, GlobalRx Prescribed fire
How did suppression go? ICS-209-PLUS, fireline engagement Suppression
How fast did a fire spread? Global Fire Atlas, FIRED, FEDS Fire behavior
What started it? GOES GLM, LIS/OTD Lightning
Who is most exposed? GlobPOP, CDC SVI, FEMA NRI Exposure & vulnerability
Which county, ecoregion, forest or GACC is this? TIGER, EPA ecoregions, PAD-US, NIFC GACC Boundaries
Training a model? WildfireSpreadTS, Sen2Fire Benchmarks

Trade-offs to check

  • Latency. Geostationary data arrives in minutes. Polar-orbiting products take hours. Annual products lag by months.
  • Resolution. Fine pixels miss nothing small but cover less history. Coarse products suit global trends.
  • Burn definition. Perimeters, burned-area pixels and burn severity measure different things. Do not mix them without checking.
  • License. Check terms before redistributing. Entries show the license the provider states.