Prediction of forest fire propagation

Application of High Performance Computing (HPC) to predict the propagation of forest fires. Integration of parallel forest fire propagation and meteorological models that are tightly coupled to improve the quality of the predictions. Use of a two-stage methodology to estimate the parameters of the local conditions of the fire with genetic algorithms and decision trees.

Predicting the fire propagation is a key issue to fight against forest fires and mitigate, or even prevent, their effects. But this prediction must be carried out under very strict time conditions. Command centers must know the expected evolution of the fire well in advance to be able to manage the field means to use them in the most efficient way

Parallel wildfire propagation simulator, based on FARSITE simulator that can be run in computing accelerators (GPUs).

Environmental systems, agriculture, urban planning

risk management services providers, farming