A bayesian network approach to assessing wildfire consequences

Wildfire risk is a function of hazard occurrence probability and potential consequences.Wildfire consequences in turn are a function of the vulnerability and exposure of the biotic and abiotic systems affected by the fire. These include among others human lives, property, infrastructure, biodiversity, soil and air quality, and agriculture production. This study focuses on the development of a wildfire building damage consequences assessment system at the meso scale, i.e. at a 1 km2 resolution. We achieve this with the construction of a Bayesian network (BN), due to its ability to facilitate the explicit modeling of the relevant parameters, their causal relationships and the associated uncertainties. Probability distributions are obtained both from observations and literature (e.g. for property values) and expert knowledge (e.g. for fire suppression performance). Numerical investigations are made with spatial datasets for the Mediterranean island of Cyprus. Results of the estimated building damage cost for given fire type are shown in maps. The presented model can be attached to a wildfire hazard model to determine wildfire risk in a spatially explicit manner.