Automation of Forest Fire Detection and Burnt Area Assessment using integrated GIS with Advanced Spatial Data Sciences for the Rudraprayag Forest Division, Uttarakhand

Abstract

Forest fire is one of the hazards that disrupts the forests and need to be monitored and managed to have a productive forest ecosystem. Forest fires pose threat to the flora and fauna and may upset the biodiversity of a region. The near real-time detection of fires using satellite-based observations has been useful in forest fire monitoring and burnt area assessment. The authors present the application of GIS environment and automation of the entire process for the near real-time forest fire detection and mapping of burnt area. The assessment was done for the Rudraprayag forest division of Mandakini Valley in the Rudraprayag district of Uttarakhand, India. The region has rich flora and fauna and faces forest fires every year leading to colossal loss. Freely available Landsat 8 and Sentinel 2 Imagery were used for the fire detection and burnt area mapping. Results were compared with the fire alerts shared by the Forest Survey of India (FSI). As FSI considers MODIS, SNPP/VIIRS data with resolutions 1 km and 375 m respectively, a large deviation was observed from the actual fire locations. Advance data science tools provide an opportunity to perform fire detection and burnt area assessment with improved accuracy. The authors successfully demonstrated here the application of advance data analytics to perform fire detection and burnt area mapping using GIS environment in an automated way. The protocol developed and adopted here can be used for other study regions of the globe to monitor and manage forest fires.