Forest Fire Severity Mapping Using Geospatial Techniques: A Case Study of a Part of Bandipur Reserve Forest, India

The changes in climate directly affect fire frequency and severity, which are likely to critically have an effect on the attribute of forest ecosystem. The present study aims to investigate the severity mapping of forest fire occurred in the year 2019 in Bandipur Reserve forest, India, using geospatial techniques. In the present study, distinct band ratios like Normalized Burn Ratio (NBR), Normalized difference Vegetation Index (NDVI) were obtained using Landsat 8 OLI images of bi-temporal data (pre-fire/post-fire). Using bi-temporal data, Difference of Normalized Burn Ratio (dNBR) and Normalized Difference Vegetation Index was obtained by subtracting the Pre/post-fire data to get the burn severity. Initially, the displacement of pixels in the burned and unburned area in the pre/post-fire NIR-SWIR and NIR-R was studied, the capacity of the two indices pre- and post-fire (bi-temporal) and post-fire (uni-temporal) was studied to analyze which one is more sensitive for severity levels discrimination. Based on the outcome, it was taken into consideration that the most appropriate way to study forest fire severity by index classification, and it was to differentiate between the pixels with unburned and burned with respect to NBR pre/post-fire difference value (dNBR). Further, both the image indices were reclassified to distinguish the pixels with high, moderate, low and very low severity. The two different sets of raster usually contain some numerical value. These two raster are overlaid and mathematically merged together to give new single output layer. The burned area were analyzed using MODIS fire points, helped in the locating hotspot area in Bandipur Reserve. This kind of study can avail valuable knowledge to prevent and monitor forest fires, and to understand the response of forest ecosystem.