We’re working to advance flood forecasting, particularly in the face of flash flooding, a leading cause of natural disaster-related fatalities. Our efforts involve investigating flood prediction systems that encompass wireless hardware, data architectures, sensor placement algorithms, state estimation techniques, and the integration of high-resolution radar data into meteorological models. We collaborate with a diverse group of experts spanning hydrology, meteorology, computer science, and crowdsourcing.

In pursuit of this goal, we deployed a network of 40 sensor nodes in the Dallas-Fort Worth region, an area that has witnessed a rise in flooding. Supported by the NSF and local authorities, our network has the potential to serve as a model for future real-time flood warning systems throughout the United States. 

Most flood prediction models require lots of expert time or computational power. We’ve developed a cheap and automated method to generate predictions for any of our sensors given only their location. 

Fundamental Advances: Analytics for flood nowcasting and forecasting, automated and computationally cheap rainfall-runoff models

Impacts: Enhance the safety and resilience of urban communities by providing timely and accurate flood forecasts, ultimately reducing the impact of flash flooding events.

Communities: Texas, Michigan