Modelling the probability of sustained flaming: predictive value of fire weather index components compared with observations of site weather and fuel moisture conditions

We investigated the likelihood that short-duration sustained flaming would develop in forest ground fuels that had direct contact with a small and short-lived flame source. Data from 1027 small-scale experimental test fires conducted in field trials at six sites in British Columbia and the North-West Territories between 1958 and 1961 were used to develop logistic regression models for ten fuel categories that represent unique combinations of forest cover, ground fuel type, and in some cases, season. Separate models were developed using two subsets of independent variables: (1) weather variables and fuel moisture measurements taken at the site of the test fire; and (2) Canadian Fire Weather Index (FWI) system components calculated from weather observations recorded at a nearby station. Results indicated that models developed with FWI system components were as effective as models developed with site variables at predicting the probability of short-duration sustained flaming in most fuel categories. FWI system components were not useful for predicting sustained flaming in spring grass fuels and had limited usefulness for modelling the probability of sustained flaming in aspen leaf ground fuels during summer conditions. For all other fuel categories, FWI system components were highly effective substitutes for site variables for modelling the probability of sustained flaming.