Modelling and Forecasting the Evolution of Forest Fire Using Vector Autoregression Processes

Abstract

The joint sample of daily values of the number of forest fire seats and the complex meteorological
index in Irkutsk region, seasons 1969-1988, is investigated. It appears that the
evolution of forest fire is well described by a vector autoregression process based model. Computer
algorithm of simulation of the future numbers of fire seats is presented. It is shown how
the algorithm can be used for reliable forecasting of the evolution of forest fire up to 2 days
ahead.