Assessment of Forest Fire Impacts on Human Health and Natural Resources using Remote Sensing

Abstract

As forest fires become more common, they are becoming a growing issue to ecosystem stability, deterioration of natural resources, and human health due to loss of vegetation, degradation of land, thermal alteration and air pollution. This work introduces a remote-sensing approach for quantifying environmental and health effects of forest fires based on multi-source satellite, air quality and meteorological observations. Reference fire data and burned area data from the MODIS and Landsat-8 and 9 datasets are used to facilitate reliable spatial and temporal assessment with sentinel-2 imagery. The image processing steps consist of atmospheric correction, cloud masking, spatial harmonization and combining multisource data. The spectral properties are used to detect the occurrence of forest-fires and burned areas and the Normalised Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) are used to measure vegetation damage. Normalized Burn Ratio (NBR) and differenced NBR (dNBR) describe burn severity; Land Surface Temperature (LST) provides a description of thermal changes related to a fire event. Observations of air quality are used to assess potential exposure of humans to fire-related pollution. The measures of accuracy, precision, recall, F1 score, intersection over union are used for evaluating performance. The framework facilitates the virtual integration of burned area mapping, vegetation degradation, fire severity, resource loss, environmental conditions of relevance to human health, as well as the assessment of ecological restoration, disaster management, and resource-conservation planning, following fire.