Improving Digital Terrain Model Accuracy for Urban 3D Flood Modeling Using UAV LiDAR

Abstract

Rapid urbanization and climate change have intensified flood risks in low-lying urban areas such as Surabaya, Indonesia. The Sepuluh Nopember Institute of Technology (ITS) campus is highly vulnerable due to inadequate drainage and flat topography. This study demonstrates the use of Unmanned Aerial Vehicle (UAV)-mounted LiDAR to generate high-resolution Digital Elevation Models (DEM) for three-dimensional (3D) flood modeling. UAV LiDAR achieved high accuracy with CE90 of 0.0387 m and LE90 of 0.0779 m, exceeding Indonesian national mapping standards. Flood simulations using HEC-RAS for 2, 5, and 10 year rainfall return periods revealed that higher DEM resolution produced more precise flood extent and depth estimates, especially in microtopographic features such as drainage channels. The novelty of this study lies in integrating UAV LiDAR-derived DEM with 3D visualization to assess flood hazards at a campus scale. While results are promising, limitations include reliance on rainfall-based scenarios and a single study site without real-Time validation. Practically, the findings support data-driven flood mitigation, drainage infrastructure planning, and the potential development of early warning systems.