Carbon Footprint Estimation Using Remote Sensing in the World’s Top Six Climate Vulnerable Cities

Abstract

Climate change and global warming pose a serious threat to urban areas globally. The rapid expansion of urban settlement by replacing green areas, forests, barren land and water bodies have increased the carbon emission and hinder environmental sustainability. This study aims to identify the land-specific carbon emission due to changing land use patterns using the Normalized Difference Vegetation Index (NDVI) of the top 6 climate-vulnerable cities of the World, which are Jakarta, Lahore, Delhi, Karachi, Lagos and Muscat. Landsat-8 satellite imagery from 2014 to 2023 with 2-year intervals was used to derive NDVI using remote sensing and Geographic Information System (GIS) approaches. The Artificial Neural Network Multi-Layer Perception Markov Chain model was utilized to predict the future NDVI patterns for the next decade. These NDVI maps are then used to generate carbon emission/absorption trends in the designated study areas. During the last decade, unplanned urbanization has increased buildup area, with Delhi experiencing the most significant urban expansion, followed by Muscat, Karachi, Lagos, Lahore, and Jakarta. There has been a noticeable rise in carbon emissions in these areas, with Karachi leading the increase, followed by Lagos, Delhi, Lahore, Jakarta, and Muscat. Conversely, the capacity to absorb carbon has declined, with Lagos experiencing the most reduction, followed by Delhi, Lahore, Karachi, Jakarta, and Muscat. Overall validation accuracy of more than 80% was shown in all the classified NDVI maps. The simulation shows that by 2030, the buildup area will increase above 80% in all selected cities, thus increasing carbon emissions with a validation accuracy of above 75%. The results were validated by comparing estimated carbon emissions with an overall carbon footprint derived from energy consumption data for every building in Sector I-14, Islamabad, segmented using the Segment Any- xxv i ABSTRACT thing Model. 77.88% of the buildup area and 231.31 tons of carbon emissions were estimated using NDVI for the year 2023, and the electricity-based carbon footprint of the whole sector is 264.84 tons, or 0.1547 tons per house per year, as validated through literature. The comparison indicates that there is an estimated error of 33.52 tons of CO2 and an accuracy of 87.45%. These results highlight the importance of remote sensing methods as a beneficial tool for policymakers and urban planners in ensuring sustainable development, preserving the city’s natural resources, and reducing the impact of changing land use types on carbon emissions.