Integrating Meteorological Dynamic Data and Historical Data into a Stochastic Model for Predicting Forest Fires Risk Maps

摘要

This paper couples a dynamic model of meteorological risk of forest fires with historical fire data in a stochastic model in order to predict forest fire risk maps. Daily Severity Rating (DSR), a meteorological risk of forest fire index, from, the Canadian Forest Fire Weather Index System (CFFWIS), results from the transformation of daily weather observations into relatively simple indices that can be used to predict fire occurrence, behaviour and impact. CFFWIS uses the daily weather observations or forecasts to calculate moisture of several fuel types and size classes, and combines them into indices of fire danger related to fire potential rate of spread, heat release, and fireline intensity. The DSR index depends only on daily measurements of air temperature (°C), relative humidity (%), 10 m open wind speed (km/h) and 24 h accumulated precip-itation (mm). DSR is extremely important for forest fire risk assessment but it is restricted to climatic factors. DSR itself is an incomplete measure of seasonal fire activity because the latter is also dependent on the ignition pattern and the available control resources. Durao proposed one Bayesian approach to calculate the local conditional proba-bilities of a forest fire occurring at any location x, given the class R(x) of predicted DSR for same location x. Suppose an indicator variable 1(x) that takes the value if a fire occurred in x, otherwise 1(x) = 0. Let us call R(x) as the classes of DSR predicted for control points and inferred by simulation for any location x. In this paper, we calculate the probability of a forest fire occurring in x, given R(x) and the historical data of fires occurrence in x, D(x): Prob {I(x)| R(x), D(x)} Both conditional probabilities Prob 1 {(x)I R(x)} and Prob {I(x)ID(.)} can be inferred at any location x. Hence conditional probability can be calculated with the method of Journel called tau model. Risk maps of forest fires can be driven from these conditional probabilities.