Spatial Modeling of Flood Vulnerability Using Google Earth Engine (Case Study: Pasar Minggu, Pancoran and Tebet Districts)

Flooding is a common problem in some parts of Indonesia, especially in urban areas with high population density. South Jakarta, especially in Pasar Minggu, Pancoran and Tebet Districts, has become one of the flood-prone areas. The main contributing factors include high rainfall, residential areas near rivers, lack of water absorption zones, and littering practices. To identify areas with the highest flood vulnerability, remote sensing using Google Earth Engine (GEE) has been implemented. This technique allows mapping flood-prone areas through satellite image analysis. Based on research, Tebet Regency has an average high level of flood vulnerability, while Pasar Minggu and Pancoran are included in the medium category. The accuracy of the analysis obtained is quite good, with a Kappa Accuracy of 83%, an Overall Accuracy of 91%, and an AUC of 0.732. This research contributes to developing a WebGIS-based flood vulnerability mapping system with direct integration of spatial modelling results from Google Earth Engine (GEE), which is still rarely applied applicatively in dense urban areas such as South Jakarta. The study also proposes a combination of new spatial weights with rigorous accuracy validation using Kappa and AUC values.