IMPACT OF FOREST FIRE ON ATMOSPHERE THROUGH REMOTE SENSING

Abstract

Forest fire has been regarded as one of the major reasons for the loss of biodiversity and degradation of environment. Forest fire causes burning of biomass which results in emission of various toxic gases along with greenhouse gases, which deteriorates air quality hence has impact on health of human being and environment. This study aims to study about forest fires and its impact on atmosphere using remote sensing and it has been portrayed through estimation the emissions from forest fires. Hence the objective of this study includes Detection of fire location, burn area estimation, estimation of aboveground biomass and estimation of emissions using space remote sensing. Fire locations can be detected using thermal infrared bands from coarse resolution sensors such as MODIS and VIIRS. There are various methods of burnt area estimation like by using spectral vegetation indices, through classification of images and through active fire data products and no. of pixels identified. Burnt area provides area burnt due to forest fire and it is used as direct input in estimation of emissions due to forest fires. Forest fire emissions in this study has been estimated based on Burning biomass method, it requires parameters like burnt area, fuel load, combustion efficiency and compound emission factor. For detection of active fire locations, fire hotspot locations have been extracted from NASA FIRMS (Fire information for resource management system) for MODIS and VIIRS sensors. Monthly fire location maps are created from 2016 to 2020 to visualise the spatial distribution of forest fire hotspots. For the assessment of Burn area, NBR (Normalized burn ratio) was calculated for Landsat 8 imagery and based on pre-fire and post-fire values, dNBR was calculated. Based on the USGS classification scheme, on the basis of dNBR values burn severity classes were identified and areas were calculated. Based on visual inspection through google earth, moderate-high severity and high severity classes have been identifies as burnt area. Most Total burn area was found in 2016. Combustion efficiency has been derived from literature review based on the forest type and the place of study. Biomass has been predicted using vegetation indices as predictor variables and reference aboveground biomass as dependable variable in Random Forest regression algorithm. Biomass for five years from 2016-2020 have been predicted by applying random forest regression analysis. For creation of regression models, Scikit library has been used. For validation of predicted biomass maps, the predicted biomass and reference biomass map for the year 2017 has been compared. Emission factors have been taken from IPCC (intergovernmental panel on Climate Change) guidelines. Emissions have been calculated by applying all these factors and its spatial distribution have been plotted on map.