Mapping Forest Fire-affected Areas Using Advanced Machine Learning Techniques in Damoh District of Central India

Abstract

Abstract

Forest fires can cause calamitous damage to the forest ecosystems and may affect climatic parameters, such as temperature, evapotranspiration, and precipitation by catalyzing changes in the local ecology of the region. The main research objective of this work was to quantify the different indices as well as to identify the changes in years 2011 and 2020 in pre- and post-fire, these parameters, such as Normalized Difference Vegetation Index (NDVI), Land Surface Temperature (LST) and Normalized Burn Ratio Index (NBR), associated with the severity of forest fires in the Damoh district of Madhya Pradesh. An attempt has been also made to understand the interrelationship of these parameters to gauge how these may correlate to determine the sensitivity to forest fires. In this context, we have incorporated advanced GIS methods for the identification of the pre-post fires for 2011 and 2020 year from Landsat 5 and Landsat 8 OLI/TIRS level -1 and Random Forest (RF) model in the Google Earth Engine (GEE) platform. Land Use Land Cover (LULC) map was categorized into five classes based on the satellite data sets. Our findings indicate elevated Land Surface Temperature (LST) values in the Northern and Central regions of the study area, reaching 32.0°C before the fire event. Subsequently, following the fire incident in the year 2011, LST escalated to 39.0°C. Similarly, in the southern and south-eastern regions of the Damoh district, LST peaked at approximately 43.0°C coinciding with the onset of a forest fire in 2020. Furthermore, our analysis revealed a negative correlation between the Normalized Difference Vegetation Index (NDVI) and LST, whereas the Normalized Burn Ratio (NBR) displayed a positive correlation with LST. These results underscore the impact of LST on forest vegetation dynamics, with LST nearing 39.0°C indicating an increased risk of forest fires. The results of this study can be used by local administration to devise an efficient policy related to forest fire management.