Forest fire and Dispersion modellings in SIMPAC

Note

In this section, you may find the list of the possible forrest fire and dispersion modellings.

A report (in French) describing the implementation of FARSITE in the SIMPAC platform is available on request. The report details the coupling with the PSPRAY model. Further information on dispersion modelling by PSPRAY may be found on the theoretical training material (part 1) and here (part 2).

FARSITE

The FARSITE model of fire growth follows a so-called vectorial or undulatory approach (Huygens principle). The fire front propagates as a continuously expanding fire polygon. The fire polygon is defined by a set of two-dimensional vertices. The number of vertices increases as the fire develops over time (the polygon expands). At regular time intervals, the expansion of the fire polygon is determined by calculating the velocity and direction of propagation from each vertex and multiplying this velocity by the duration of the time step. The direction and velocity of propagation normal to the fire front is determined from the direction and velocity of maximum propagation by an elliptical transformation. The use of an assumed fire shape, in the case of FARSITE an ellipse, is necessary because only the propagation speed of the leading portion of a fire is predicted by the fire propagation model. Fire spread in all other directions is inferred from the forward spread rate using the mathematical properties of the ellipse.

_images/farsite_pricipe.png

Illustration of the Huygens principle using an elliptical wavefront. (A) Uniform conditions use wavelets of constant shape and size to maintain the elliptical shape of the fire over time. (B) Non-uniform conditions showing the dependence of the wavelet size on the local fuel type, but also the dependence of the shape and orientation of the wavelets on the local wind slope vector.

For the SIMPAC platform, forest fire growth and pollutant dispersion takes into account the following points:

  • The meteorological conditions result from the chosen PSWIFT simulation.

  • FARSITE outputs contain the flame perimeter (shape format) at each time step and the reaction intensity (kW/m2) (raster format) at each grid point. The model estimates the fire propagation. It does not describe the burning surfaces, but only the progress of the fire front.

  • For the modeling of the dispersion of the smoke from the fire, we assume that only the areas close to the fire front emit non-negligible quantities of pollutant. In other words, the chosen hypothesis is that the surface that burns at a time t is the surface included between 2 time steps of the FARSITE modelling, i.e. 15 minutes.

  • The intensity of the reaction in kW/m2 is averaged over the whole burning surface at each time step. The final output is a Shape containing the different burning surfaces with their power at each time step. This power allows the PSPRAY dispersion model to estimate the overheight of the fire plume.

Important

  • In the version of FARSITE currently implemented in the SIMPAC platform, there is no modulation of the quantities emitted according to the type of fuel and its availability.

  • There is also no estimation of the species emitted according to the type of fuel. Only a tracer species is emitted, and the emission flux is assumed to be 1kg/h.

  • The coupling is partly based on the specifications chosen for the PSPRAY dispersion model. In particular, one particle per emission time step and per source is emitted, similar to the trajectory configuration.