Abstract
The extreme flood disaster in September 2025 in the Tukad Badung River Basin, Bali, had a significant impact, triggered by high rainfall reaching 385 mm per day. This study aims to analyze the distribution of flood inundation and identify its relationship with land cover conditions in the area. The methodology used is remote sensing integration using Sentinel-1 (SAR) and Sentinel-2 satellite imagery through the Google Earth Engine platform. Inundation detection was carried out using the change detection method using an adaptive threshold (
P
85
+0,3), while land cover classification used the Random Forest algorithm. The results showed that the total flood inundation area in the Tukad Badung watershed reached 356,831 hectares. The flood mapping accuracy test resulted in an Overall Accuracy (OA) value of 93.08% with a Kappa Coefficient of 0.86. Spatial analysis revealed that built-up areas was the most dominant class affected by flooding, reaching 63.4% of the total inundation area. This indicates that the massive impermeable surface in urban areas significantly reduces infiltration capacity and exacerbates surface runoff volume. These findings are expected to be a strategic reference for disaster mitigation efforts and spatial planning in Bali Province.