Machine learning based urban sprawl assessment using integrated multi-hazard and environmental-economic impact

Abstract

Abstract The increasing demand for land development due to human activities has fueled urbanization. However, uncontrolled urban development in some regions has resulted in urban environmental problems arising from an imbalance between supply and demand. This study aims to develop an integrated model for evaluating and prioritizing the management of hazardous urban sprawl in the Bandung metropolitan region of Indonesia. The novelty of this study lies in its pioneering application of long-term remote sensing data-based and machine learning techniques to formulate an urban sprawl priority index. This index is unique in its consideration of the impacts stemming from human economic activity, environmental degradation, and multi-disaster levels as integral components. The analysis of hazardous urban sprawl across three distinct time periods (1985–1993, 1993–2008, and 2008–2018) revealed that the 1993–2008 period had the highest increase in human economic activity, reaching 172,776 ha. The 1985–1993 period experienced the highest level of environmental degradation in the study area. Meanwhile, the 1993–2008 period showed the highest concentration of multi-hazard locations. The combined model of hazardous urban sprawl, incorporating the three parameters, indicated that the highest priority for intervention was on the outskirts of urban areas, specifically in West Bandung Regency, Cimahi, Bandung Regency, and East Bandung Regency. Regions with high-priority indices require greater attention from the government to mitigate the negative impacts of hazardous urban sprawl. This model, driven by the urban sprawl priority index, is envisioned to regulate urban movement in a more sustainable manner. Through the efficient monitoring of urban environments, the study seeks to guarantee the preservation of valuable natural resources while promoting sustainable urban development practices.