Canadian Forest Service Publications
Mapping wildfire susceptibility with the BURN-P3 simulation model. 2005. Parisien, M.-A.; Kafka, V.; Hirsch, K.G.; Todd, J.B.; Lavoie, S.G.; Maczek, P.D. Natural Resources Canada, Canadian Forest Service, Northern Forestry Centre, Edmonton, Alberta. Information Report NOR-X-405. 36 p.
Issued by: Northern Forestry Centre
Catalog ID: 25627
Availability: PDF (download)
To optimize strategic planning, resource management in fire-dominated ecosystems requires an understanding of the probability of wildfire occurring and spreading at different points on a landscape. This report describes an approach to evaluating wildfire susceptibility, or burn probability (BP), for fire-prone landscapes such as the boreal forest of North America. BURN-P3 (probability, prediction, and planning) is a landscape-level simulation model producing BP maps. The model combines deterministic fire growth based on the Canadian Fire Behavior Prediction System and spatial data for forest fuels and topography with probabilistic fire ignitions and spread events derived from historical fire and weather data. Model components include the location and frequency of ignitions, the rate at which fires escape initial attack and become large, the number of days on which each fire achieves significant spread, the fire weather conditions associated with these spread event days, and the deterministic fire spread. For a given landscape, BP is simulated for a single annual time step, or iteration, based on 500 to 1000 Monte Carlo simulations. A case study of the application of BURN-P3 was undertaken for a 15 × 106 ha boreal mixedwood area of central Saskatchewan. The BP values varied considerably within the study area. Regions with a high BP were highly localized (clustered distribution), largely because of the configuration and continuity of flammable forest fuels. These results highlight the importance of landscape features, such as lakes and recent burns, to wildfire susceptibility, and suggest that assessments based solely on stand-level characteristics may be inadequate.