Forest Fire Occurrence

Recent work has seen the development of a generalized linear modeling framework for fire occurrence prediction which was used to develop daily models of lightning and human-caused fires for the province of Ontario.

Ignition vs. SDMC
Relationship observed in two ecoregions in Ontario between
daily fire occurrence and FFMC

Human-caused fire occurrence

Poisson regression was used to develop models of expected daily fire occurrence for each of the ecoregions in Ontario's fire management region. One of the most important factors influencing fire occurrence was found to be the moisture content of surface litter fuels (as represented by the Fine Fuel Moisture Code (FFMC) of the Canadian Forest Fire Weather Index (FWI) System). Expected number of fires also varied with ecoregion, time of year and fire cause type.

Further work has examined the linkages between fire occurrence and human influences in forested areas, such as population level, number of communities, the presence of roads and railways. This research lead to the development of models of expected fire occurrence per length of roadway or railway in a district as well as number expected fires per year as a function of the number of communities in a district.

Lightning-caused fire occurrence

A lightning fire occurrence predition system was developed by modelling the probability of the ignition of a holdover fire by lightning as one process and the probability of arrival or detection of that holdover fire as a second process. Sheltered Duff Moisture Code (SDMC: see the description under fuel moisture projects) was found to have the strongest influence on lightning fire ignition, while the FFMC was found to be generally the best indicator of fire arrival (detection).

Ignition vs. SDMC
The observed relationship between raw ignition
probability and SDMC for eastern and western Ontario.

The fire occurrence models developed were implemented and tested operationally in the province of Ontario during the summer of 2006. The system produces daily maps of the expected location of holdover fires (after a lightning storm) and the expected location of lightning fire arrivals each day. Similar models of lightning fire occurrence are under development in Saskatchewan as part of a cooperative project with the fire management program of the Saskatchewan Environment department.