Analisis Spasial Untuk Memprediksi Risiko Bencana Banjir di Daerah Aliran Sungai

Abstrak

Land use change significantly alters the hydrological characteristics of watersheds, especially in rapidly urbanizing river basins like the Citarum, Indonesia. This study aimed to assess the impact of land cover transformation on flood risk by integrating spatial modeling using Cellular Automata–Artificial Neural Network (CA–ANN) with hydrological simulation via HEC-HMS. Multi-temporal Landsat-8 imagery (2014–2024) was used to classify and project land cover to 2029, while Curve Number (CN) values derived from land use types were employed to estimate surface runoff. The results indicated a substantial increase in built-up areas, particularly in the midstream and downstream regions, replacing agricultural lands and reducing vegetative cover. This shift significantly raised CN values across most sub-watersheds, resulting in increased peak discharge, especially in Cikeruh, Cikapundung, and Cibeet. Flood risk mapping showed that over 50% of the sub-watersheds fall into the moderate to high-risk categories, driven by impervious surface expansion and declining infiltration capacity. This integrated spatial–hydrological approach underscores the importance of adaptive land use planning and watershed-based flood mitigation strategies. The findings offer a scientific basis for ecosystem-based disaster risk reduction and inform policy-making in flood-prone urbanizing basins