OPERATIONAL FOREST FIRE-GROWTH PREDICTIONS FOR CANADA

This paper presents an operational model that predicts fire growth for wildland fires occurring in Canada. National fuels and elevation grids, forecasted weather and fires detected by remote sensing are entered into a fire growth model. Predicted fire perimeters are mapped and presented to the public over the Internet through the Canadian Wildland Fire Information System (http://cwfis.cfs.nrcan.gc.ca/). Current wildland fires are detected nationally using MODIS and NOAA/AVHRR satellite-based detection systems (burning areas detected by these systems are referred to as hotspots). Regions selected for fire-growth modelling include those with hotspots occurring near a community, and clusters of hotspots in areas of high fire danger. For each selected region, a fire-growth simulation environment is assembled. Fuel type data from several fire management agencies is available in grid format at a resolution of 100 m or less; in areas where such data is not available, a national fuels map based on SPOT VGT land cover and forest industry is used. Similarly, terrain data is available from a variety of sources. Current hotspots are used as ignition points while past hotspots are used to delineate area burned. Surface wind, temperature and dew-point values (forecasted by Environment Canada) are used to determine the fire weather conditions at the fire location. The use of this model is illustrated for a large fire in Wood Buffalo National Park, which burned nearly 200 000 ha in July 2007.