Integrating Geo-AI with RS & GIS for comprehensive assessments of urban land cover transformations and integrated responses

Abstract

Poorly conceived urbanization negatively affects the urban environment, particularly in developing nations such as Pakistan. The resultant degradations undermine urban ecological and environmental integrity. Resource, knowledge, and governance-related constraints further aggravate the situation. This study compares spatial-temporal trends in land use and land cover (LULC) transformation from divergent contextual settings such as Attock, Faisalabad, and Islamabad, Pakistan. The cities were selected for their contrasting models of urban planning. The study can be a yardstick for all similarly planned towns in developing countries. In this study, we used the Landsat 5 and 8 in the Google Earth Engine (GEE) from 1990 to 2020. We also used high-resolution imagery to validate the results in the GEE. We modified the supervised classification with geo-Ai. From 1990 to 2020, all three cities experienced significant changes in land cover. In Attock, agricultural land expanded from 65.48% to 74.51%, while barren land and tree cover decreased, and built-up areas grew substantially, reaching 13.47%. Faisalabad saw a notable increase in built-up areas from 10.45% to 26.51%, with a corresponding decline in agricultural land from 86.24% to 68.87%. Islamabad’s built-up area rose dramatically from 6.02% to 29.04%, with reductions in tree cover and agricultural land, reflecting intensified urbanization. The concluding assessments demand a focus on LULC management in countries like Pakistan to tackle hyperactive urbanization. Studies such as this are significant for possibly addressing this threat in similar contextual settings.