John Floroff
Creating a Wildfire Model from Remote Sensing of Boreal Forests in Northern Ontario © 2012
Ontario’s boreal forests cover 41,171,000 hectares. Forest fires in the boreal forest are the main natural disturbance and 2011 was a very active year. In Northern Ontario 632,533 hectares were burned compared to 14,823 hectares in 2010 and 20,656 hectares in 2009. The overall objective of this research project is to investigate the extent of the burn areas utilizing remote sensing. Remote sensing provides a cost effective method for monitoring forest disturbance such as forest fires in vast remote areas, and can contribute insight to policy and management objectives. An analysis using remote sensing techniques was undertaken to examine the extent of several forest fires that occurred during 2011 in Northern Ontario. Landsat 5 Thematic Mapper images were acquired for study areas near Wabakimi Provincial Park and Pickle Lake with the time period being from 2009-2011 for both fire study areas. The post-fire image acquisition dates were as close as possible to fire extinguishment to minimize temporal distortions. The analysis utilized the Normalized Difference Vegetation Index (NDVI), Principal Component Analysis (PCA), Tasselled Cap Transformation (TCT), Normalized Burn Ratio (NBR), unsupervised classification and image differencing operations. Favourable conditions for wildfires such as dry conditions, thunderstorms, strong winds and large amounts of fuel were the main factors contributing to the 2011 fire season and they influenced fire behaviour and progression in the study areas. A total of 586 square kilometres (17% of the study area) and 450 square kilometres (18% of the second study area) were burned as a result of the fires.