How the wildfire spread simulator works

Full transparency on the model behind the simulator: the published fire-science equations it uses, the live government data that feeds them, how it’s verified, and exactly what it does and doesn’t do.

What it is

The simulator is a real implementation of the Rothermel (1972) surface fire-spread model with Van Wagner (1977) crown-fire logic and Byram (1959) flame length — the same physics used in BehavePlus and the U.S. federal fire-behavior tools — driven by real fuel, weather, and terrain data for the location you choose. It is built to make fire behavior intuitive and is rigorously grounded, but it is an educational/illustrative tool, not an operational forecast. For active fires and evacuations, always follow Watch Duty, InciWeb, and local authorities.

You can run it three ways:

  • •Your own address — centers on your rooftop and shows the live conditions feeding the model — current wind, temperature and humidity, dead-fuel moisture, the LANDFIRE fuel model, and canopy.
  • •A famous location — recognizable wildland-urban-interface spots so the terrain is familiar.
  • •A historic fire — centers on the documented ignition point and replays the actual fire-day wind from the historical-weather archive.

The computation, step by step

From a location to a fire front, every input is real and every step is a published equation.

1

Surface fuel model — LANDFIRE

We identify the Anderson 13 fire-behavior fuel model (FBFM13) at the location from the USGS/USFS LANDFIRE 2024 ImageServer. Developed parcels read as non-burnable, so we sample concentric rings (≈0.6 km then ≈1.6 km) outward to find the surrounding wildland fuel that would actually carry fire toward a structure.

2

Dead fuel moisture — current weather

We pull current temperature and relative humidity (open-meteo) and convert them to fine dead-fuel moisture using the Simard (1968) equilibrium-moisture-content equations. 10-hour and 100-hour moistures are offset from the 1-hour value. The intensity selector shifts this dryness up or down.

3

Wind — NWS, yours, or the real fire-day

By default we use the live wind from the nearest National Weather Service station. You can override it. For a historic fire, we pull the actual fire-day wind from the open-meteo historical archive at the ignition point and date. The 20-foot wind is reduced to a midflame wind with a fuel-dependent wind adjustment factor (Albini & Baughman 1979).

4

Slope — USGS elevation

We sample a ring of elevations around the point (open-meteo / USGS) to estimate slope steepness and the uphill direction, which accelerate and steer the fire.

5

Rate of spread — Rothermel (1972)

The fuel model, moistures, midflame wind, and slope feed the full Rothermel surface fire-spread equations — reaction intensity, propagating flux ratio, wind and slope factors, and the heat sink — to compute the head rate of spread. This is the same model in BehavePlus and the federal fire-behavior tools.

6

Flame length — Byram (1959)

From the reaction intensity, residence time, and rate of spread we compute Byram’s fireline intensity and flame length: L = 0.45·I⁰·⁴⁶ (I in Btu/ft/s).

7

Crown fire — Van Wagner (1977)

Using LANDFIRE canopy base height, bulk density, and cover, we test Van Wagner’s criteria: the surface fire torches into the canopy when its intensity exceeds the critical intensity I₀ = (0.01·CBH·(460 + 25.9·FMC))^1.5, and becomes an active crown fire when crown spread exceeds the critical mass-flow rate (3.0 ÷ canopy bulk density). The result is classified Surface, Torching/passive, or Active crown fire.

8

Fire shape & embers

Wind stretches the front into an ellipse using Anderson’s (1983) length-to-breadth ratio. A downwind ember plume is drawn ahead of the front, with an estimated spotting distance that scales with fire type and wind.

Data sources

LANDFIRE 2024 (USGS/USFS)

Surface fuel model (FBFM13), canopy base height, bulk density, and cover.

National Weather Service / NOAA

Live wind speed and direction.

Open-Meteo

Current temperature & humidity (fuel moisture) and the historical wind archive for fire replays.

USGS elevation

Terrain slope and aspect.

How it’s verified

The model is checked against published values using a controlled-input endpoint that runs the equations on fixed fuel and weather. Its rate of spread reproduces the standard benchmarks — for example, fuel model 1 (short grass) at 6% dead-fuel moisture with no wind or slope computes ≈ 4.2 chains/hour, matching the published figure; closed timber litter (fuel model 8) stays near 1.6 chains/hour; chaparral (fuel model 4) runs fast with tall flames. Spread rates rank correctly across all 13 fuel models and respond correctly to wind, slope, and rising or falling fuel moisture. An independent line-by-line code review against Rothermel (1972) and Andrews (2018) confirmed the equations and constants, including the moisture, mineral, wind, and slope terms.

Assumptions & limitations

  • •It models surface fire spread with a basic crown-fire classification. It does not model the rate of an independent crown run, long-range spotting ignitions, fire-atmosphere feedback (plume-dominated fire), or suppression.
  • •Midflame wind is estimated from the 20-foot wind with a standard fuel-dependent adjustment factor, not a measured in-stand wind.
  • •Live-fuel moisture uses representative fire-season values; dead-fuel moisture is estimated from current (or archived) temperature and humidity, not field samples.
  • •The fuel model is the dominant nearby wildland fuel from a coarse (30 m) national layer; a specific parcel can differ.
  • •The spotting distance is an order-of-magnitude estimate, not a physical firebrand-transport (Albini) calculation.

References

  1. Rothermel, R.C. (1972). A mathematical model for predicting fire spread in wildland fuels. USDA Forest Service, INT-115.
  2. Albini, F.A. (1976). Estimating wildfire behavior and effects. USDA Forest Service, INT-30.
  3. Anderson, H.E. (1982). Aids to determining fuel models for estimating fire behavior. USDA Forest Service, INT-122.
  4. Anderson, H.E. (1983). Predicting wind-driven wildland fire size and shape. USDA Forest Service, INT-305.
  5. Byram, G.M. (1959). Combustion of forest fuels. In: Forest Fire: Control and Use.
  6. Van Wagner, C.E. (1977). Conditions for the start and spread of crown fire. Canadian Journal of Forest Research, 7(1).
  7. Simard, A.J. (1968). The moisture content of forest fuels. Canadian Dept. of Forestry & Rural Development.
  8. Andrews, P.L. (2018). The Rothermel surface fire spread model and associated developments: A comprehensive explanation. USDA Forest Service, RMRS-GTR-371.
Open the simulator →