Physically based dimensionless features for pluvial flood mapping with machine learning

May 30, 2025

By Lalitha Krishnamoorthy, Jinshu Li and Curtis Smith

When minutes matter, communities need fast, accurate flood forecasts to protect lives and property

Flash floods are among the deadliest natural disasters, accounting for up to 90 percent of flood-related deaths each year. Predicting these events quickly is critical—but not easy. Traditional flood modeling methods—like detailed 2D hydraulic models—are too slow and resource-intensive to meet real-time needs.

In collaboration with The Water Institute, our team recently published a study in Water Resources Research. It combines the power of physics with the speed of machine learning to create flood maps that are faster and more adaptable across diverse landscapes. This breakthrough could transform how cities, engineers, and emergency managers prepare for and respond to flash floods.

Read Study

As our artificial intelligence and digital lead, Lalitha is set to lead technology strategy and standardize our digital platforms. She co-founded OpenTeams Global and has a doctorate in neuro-symbolic artificial intelligence.
Contact Lalitha 
Jinshu is a data scientist who addresses complex hydrological and water resource challenges through statistical methodologies, machine learning, and mathematical modeling.
Contact Jinshu 
A senior associate, team leader, and technical expert, Curtis works on flood risk identification. He applies digital solutions such as cloud computing, statistics, and watershed modelling to maximize value for the communities he supports.
Contact Curtis 