Creating a forest fire database for the Far East of Asia using NOAA/AVHRR observation

Abstract

This study proposed a method for burned area accounting that uses data from the Advanced Very High Resolution Radiometer (AVHRR) onboard the National Oceanic and Atmospheric Administration (NOAA) satellite series. As an area of interest, the boreal forest zone located in the Far East region of Asia was used. The burn scar mapping algorithm consists of two parts. The first is a multi‐channel threshold algorithm used for detection of real‐time burning spots in the boreal environment. The second part uses an abrupt change‐detection technique in Normalized Difference Vegetation Index (NDVI) in an 18‐year NDVI time series. Both parts of the algorithm are connected together in a complementary manner, and a forest burn scar mask is obtained for each month and consequently for each year from 1984 to 2001. The validation of the dataset was performed using data from the literature, forestry management organizations and the Landsat Thematic Mapper (TM). The comparison between those validation data and our forest fire dataset shows a satisfactory level of agreement. If the forest fire history is required for other regions in the boreal zone, the proposed methodology could be extended to that region given that a sufficient data archive is available.

Acknowledgments

We would like to thank Sumio Tamba at the Hirosaki University in Hirosaki for advice on the PaNDA software, Toshihiro Nemoto at the Institute of Industrial Science in Tokyo for managing the AVHRR archive and Ochi Shiro at the Institute of Industrial Science in Tokyo for help with computer hardware.