An improved cellular automaton model for simulating fire in a spatially heterogeneous Savanna system

Developments in and around game reserves and ranches in South Africa have led to controlled burning becoming a necessary and regular activity. The management objectives of these fires are well-defined, and thus predicting the duration and extent of a burn is vitally important. Testing scenarios via computer simulation is desirable since this removes the potential risks associated with fire, whilst at the same time ensuring that management policies are attained. There are various approaches to developing a spatial simulation fire model. In this article we present a cellular automaton (CA) model that is capable of predicting fire spread in spatially heterogeneous Savanna systems. The physical basis of Rothermel's fire spread model (1972) was modified to a spatial context and used to improve the CA model introduced by Karafyllidis and Thanailakis (1997). The proposed fire model was verified using data for three human-induced fires in the Mkuze Game Reserve, South Africa, and was found to satisfactorily predict spatial fire behaviour.