Dinamika Spasiotemporal NDVI, NDBI, dan LST Kota Denpasar 2013–2025: Tren Robust Piksel dan Hubungan Spasial-Temporal Berbasis Landsat 8

Abstract

Changes in vegetation cover and built-up surfaces influence urban thermal characteristics; however, their temporal patterns in Denpasar City have not been consistently examined on an annual basis. This study analyzed the temporal and spatial trends of the Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), and Land Surface Temperature (LST), identified pixel-level change directions, and assessed relationships among the variables during 2013–2025. A total of 435 Landsat 8 Collection 2 Tier 1 Level-2 images were processed in Google Earth Engine to generate 13 annual median composites at 30 m spatial resolution. The analyses included annual statistics, linear regression, pixel-level trend-slope classification, and Pearson correlation using 1,840,845 pixel-year pairs. Median NDVI decreased significantly by 0.00558 per year (R² = 0.634; p = 0.001), while median NDBI increased significantly by 0.00420 per year (R² = 0.582; p = 0.002). Overall, 52.60% of pixels showed declining NDVI and 55.41% showed increasing NDBI. Median LST showed no significant linear trend (−0.173 °C per year; R² = 0.199; p = 0.127), although 81.76% of pixels had negative slopes, 14.04% were stable, and 4.20% increased. Correlations were strongly negative for NDVI–NDBI (r = −0.768), moderately positive for NDBI–LST (r = 0.602), and weakly negative for NDVI–LST (r = −0.325). These findings indicate widespread vegetation decline and built-up expansion, while thermal changes were more strongly affected by interannual fluctuations and local variability. Areas with declining NDVI, increasing NDBI, and rising LST should be prioritized for green-space evaluation and land-use control.