Evaluation of ERS SAR data for prediction of fire danger in a Boreal region

Abstract

Research was conducted to determine the utility of Synthetic Aperture Radar (SAR) data for measuring the fuel moisture status of boreal forests as reflected in Fire Weather Index Codes. Three years (May to August 1992–1995) of SAR data from the European Remote Sensing Satellite (ERS) were analysed over the 1990 Tok Alaska burned and adjacent unburned black spruce forests. Corresponding Fire Weather Index Codes of the Canadian Forest Fire Danger Rating System were obtained from Tok Area Forestry, Station number 500720. Strong relationships were expected between the SAR data and fire codes because of the dependence of ERS SAR backscatter on the moisture status of forests and exposed surfaces (burn scars). Astepwise multilinear regression procedure was used to analyse the relationships. Three statistically significant multilinear regression models resulted from this analysis procedure. The models developed show there is potential for using ERS SAR backscatter to generate indicators that are related to Fire Weather Index, Duff Moisture Code, and Drought Code. This research could lead to the ability for remote prediction of fire danger over large regions at relatively fine spatial resolution with minimal weather information.