The fire localization using a distributed consensus finding algorithm in a wireless sensor network is described. The fire is circumscribed by a circle. The information is available at all sensor nodes that are alive, which makes it robust against failures and losses. Minimizing energy consumption is crucial for sensor nodes that have to function autonomously as long as possible. Therefore, the speed of convergence of the consensus finding algorithm has to be optimized. We argue that optimizing, as is customary, the asymptotic speed of convergence is not the best method when a consensus value of low precision is su±cient.