Development and Application of an Urban Flood Risk Assessment Method under Climate Change with an Exploration of AI-Assisted Applications

This study addresses the challenges of urban flooding induced by climate change by proposing a refined flood risk assessment methodology to provide scientific support for the formulation of flood adaptation strategies. Focusing on the unique disaster characteristics of Taichung City, the research integrates AR5 and AR6 rainfall scenario data provided by Taiwan’s National Science and Technology Center for Disaster Reduction. Utilizing the physiographic drainage-inundation model (PhD model), the study simulates flood depth and distribution characteristics under varying rainfall intensities, complemented by historical data and local intelligence for model calibration. This approach enables precise identification of high-risk areas and systematically characterizes flood process, offering a quantitative foundation for planning flood control infrastructure and adaptation strategies. The results address the lack of quantitative data in current urban flood risk assessments and establish a reference framework for scientific risk evaluation under extreme climate scenarios.

For extended applications, the study explores the potential of integrating flood risk information with artificial intelligence (AI) technology, specifically through the development of an intelligent water level recognition model. This model leverages existing CCTV systems for water level monitoring, employing simulated imagery for training and validation. It demonstrates potential for real-time water level monitoring and flood early warning capabilities. While further optimization and field testing are necessary, this approach holds promise for enhancing disaster mitigation and emergency response efficiency, providing valuable insights for addressing future challenges posed by extreme climate conditions.