Article
Pemodelan Spasial Ancaman Banjir pada Lahan Sawah Menggunakan Google Earth Engine dan Citra Sentinel
Jurnal Penelitian Pendidikan Geografi, Vol. 11 No. 2 (2026), hlm. 287–298
Abstrak
Rice productivity in Indonesia is highly vulnerable to hydrometeorological disasters such as flooding, which can damage crops and threaten national food security. Therefore, efficient and rapid spatial monitoring based on satellite data is urgently needed. This study develops a spatial modeling approach for flood hazards on rice fields in Bone Bolango Regency and Gorontalo City using the Google Earth Engine (GEE) platform. The methodology integrates Sentinel-2 optical imagery via the Random Forest algorithm for rice field classification and Sentinel-1 radar imagery through SAR Backscatter Change Detection to map multi-temporal flooding. The analysis identified a total of 394.81 hectares of flood-prone areas (243.98 ha in Bone Bolango and 150.83 ha in Gorontalo City). Confusion matrix evaluation showed a rice field classification accuracy of 92% (Kappa 0.77) and a flood model Overall Accuracy of 88% (Kappa 0.82), with user and producer accuracies for the high-hazard class exceeding 91%. This cloud-computing-based integration of radar and optical imagery proves highly effective. These findings are crucial for developing geospatial monitoring systems to support disaster mitigation planning and sustainable agriculture.