Probability of Sustained Ignition

Bitter Creek test fire
Picture of a test fire from the Bittern Creek, Saskatchewan site.
(From J. Beverly (CFS) collection of test fire photos)

Re-analysis of the Canadian Forest Service small scale test fire database

This project is being carried out in collaboration with Dr. Jen Beverly of the Canadian Forest Service (CFS) Northern Forestry Centre in Edmonton Alberta. From 1940 to 1961 the CFS ran an extensive experimental field program where approximately 20,000 small test fires were ignited in a number of different stand types. Observations of the resultant fire behaviour, fuel moisture and ambient weather conditions were recorded. These observations formed the basis of numerous fire hazard tables and fire danger rating systems across Canada, which eventually became the Canadian Forest Fire Weather Index System.

We have begun a re-analysis of this dataset to develop models of the probability of sustained flaming for a range of different forest types (e.g., pine, spruce, aspen, mixedwood, balsam fir, Douglas fir) and forest floor cover material (e.g., grass, lichen, feathermoss, needle, leaf). Probability of sustained flaming has been found to be well correlated with expected number of people caused fires.

This historical dataset also contains information such as fuel moisture for several different fuel strata (e.g., litter, duff, moss, lichen, woody material), flame length, depth of burn, and vigour of spread (the latter is used to determine a binary variable that indicates whether or not the ignited fire could sustainably spread).