A new technique for visualization of forest fire smoke plumes using MODIS data

Forest fire smoke detection by satellites is important and required for monitoring air pollution and human health. MODIS smoke detection algorithms are under development. The common problem is to separate smoke from clouds. To overcome this issue we propose a new visualization technique of a false-color composite image that composed of Smoke Reflectance Index (SARI), MODIS channel 7 reflectance, and Water Index (WI). The SARI and WI were developed in this study. The false-color composite image shows smoke in reddish and clouds in pink-white. Smoke pixels are easily identified and sampled. Overall smoke pixels are detected by their training dataset. In this paper, we present a case study of Russia and Mongolian forest fire in 2009. The result of smoke detection was compared to those of existing method. It was confirmed that the proposed method detected smoke pixels more accurately.