AI flood map reveals 11 million Americans may be omitted from official risk zones
Official flood maps shape disaster preparedness, insurance decisions and urban planning, but large parts of the United States remain unmapped or under-mapped. This means some communities may be left unaware of the risks they face, limiting their ability to prepare for future floods.
A deep learning framework developed by researchers from the National University of Singapore and Tsinghua University has revealed that 11 million Americans may be omitted from official flood risk zones. The AI model, which analyzed existing flood records and terrain data, created a spatially complete 30-meter flood hazard map for the contiguous United States.
This new map indicates that the scale of flood exposure is much larger than previously recognized, with gaps affecting socially vulnerable populations such as the elderly and children. The model offers a scalable and cost-effective alternative to traditional flood mapping methods, which can be expensive and time-consuming. While the AI-generated maps are not intended to replace official regulatory flood maps, they serve as a valuable public guide to highlight overlooked risks and support more targeted resilience planning.
Written by urgent.news from Phys.org's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
