Analysis of Spatiotemporal Dynamics and Associated Factors of Demolition-Led Urban Renewal in Shanghai Based on Time Series Sentinel-2 Imagery

Urban renewal is central to stock optimization in megacities, yet limited long-term monitoring data constrain understanding of its spatiotemporal dynamics and associated factors. We developed a Sentinel-2 framework combining the Land Clearing Index, vegetation and water indices, and prior built-up extents to map demolition-led urban renewal without supervised classifier training samples. The framework generated nine 10-m interval maps for Shanghai labelled 2016–2024, achieving mean annual precision, recall, and F1-score of 85.65%, 85.27%, and 85.34%, respectively. Combined with socioeconomic and spatial data, these maps revealed pronounced temporal fluctuations and spatial heterogeneity. Temporally, renewal activity fluctuated alongside policy milestones and was positively associated with secondary-industry share, tertiary-industry growth, and household size. Spatially, hotspots shifted from northern to southeastern Shanghai, while renewal intensity formed a centre–periphery gradient centred on Putuo District. At the district scale, renewal intensity was higher in districts with larger registered elderly populations, greater building density, and better rail-transit accessibility. At the grid scale, model predictions of renewal were associated with population count, road density, and proximity to metro stations, parking facilities, and residential areas. These spatial patterns were consistent with the rent-gap hypothesis. This study provides a framework for long-term monitoring of demolition-led urban renewal and insights into its spatiotemporal dynamics and associated factors in megacities.