Mitigating wildland fire hazard using complex network centrality measures

We show how to distribute firebreaks in heterogeneous forest landscapes in the presence of strong wind using complex network centrality measures. The proposed framework is essentially a two-tire one: at the inner part a state-of- the-art Cellular Automata model is used to compute the weights of the underlying lattice network while at the outer part the allocation of the fire breaks is scheduled in terms of a hierarchy of centralities which influence the most the spread of fire. For illustration purposes we applied the proposed framework to a real-case wildfire that broke up in Spetses Island, Greece in 1990. We evaluate the scheme against the benchmark of random allocation of firebreaks under the weather conditions of the real incident i.e. in the presence of relatively strong winds.