Satellite Imagery-Based AHP Model to Identify the Spatial Heterogeneity of Fire Incidents in Dhaka City Corporation, Bangladesh

Abstract

Fire is one of the severe hazards in developing and developed countries, threatening life, structures, property, and the environment. Hundreds of fire incidents occur every year in Bangladesh, causing massive damage. The study identified the spatial distribution of fire incidents in Dhaka City Corporation using satellite imagery and geospatial analysis, which was done through Google Earth Pro (x, y coordinate) and ArcGIS. A dataset is developed, listing fire incident location, time, reason, casualty, and land use from various sources, such as fire stations and newspapers. High-resolution satellite imagery available in Google Earth Pro, sourced from Maxar Technologies and visual representations of Landsat 8 and Sentinel data, was utilized to delineate land cover types around fire incident sites. Manual digitization was performed in ArcMap 10.8, focusing on visually distinguishable features to classify land cover around fire incident sites. The land covers include settlements, vegetation, water bodies, open fields, and roads. Correlation analysis was done using SPSS software to determine the relationship between the fire incident number of settlement density and the fire incident number of settlement type. The AHP model has been applied to measure the weight of different land cover classes according to fire incidents. From the correlation analysis, it is identified that residential buildings are the most vulnerable to fire hazards. The ranking of other settlement types in order of vulnerability includes slums, commercial buildings, service buildings, and industrial buildings in descending order. Therefore, special precautions and actions by the authority and policymakers against fire hazards in residential buildings and slums need to be taken.