Mapping of Urban Land Surface Temperature Trends in Makassar (2000–2020) Using a Harmonic Model for Continuous Change Detection and Classification

Abstract

Rapid urban expansion modifies land cover, alters vegetation dynamics, and intensifies urban surface warming. This study examines long-term vegetation change, urban expansion, and land surface temperature (LST) dynamics in Makassar City, Indonesia, from 2000 to 2020 using multi-temporal Landsat imagery. Vegetation conditions were quantified using the Normalized Difference Vegetation Index (NDVI) derived from annual cloud-free composites. Temporal vegetation trajectories were analyzed using continuous change detection and classification (CCDC) based on a harmonic regression model, while urban expansion was identified through persistent vegetation loss. LST was retrieved using a consistent, sensor-specific workflow to enable interannual comparison. The results reveal a pronounced decline in vegetation between 2010 and 2015, with the strongest reductions occurring in areas where NDVI values decreased below 0.1, indicating conversion from vegetated surfaces to impervious urban land cover. During this period, urbanized areas expanded by approximately 1,500–2,000 hectares. Concurrently, mean LST increased from 22.21 °C in 2000 to 28.16 °C in 2020, representing an increase of 5.95 °C in mean urban land surface temperature (LST), primarily driven by vegetation loss and the expansion of impervious surfaces. Spatial analysis shows that the most substantial LST increases coincide with zones of sustained vegetation loss and urban expansion. These findings demonstrate the effectiveness of long-term satellite time-series analysis for detecting gradual and abrupt urban-induced environmental changes. The proposed framework provides robust evidence for monitoring vegetation degradation and urban surface warming over extended periods, offering valuable insights for urban planning and climate adaptation strategies in rapidly developing cities.