Predicting flooding effects with AI

Mott MacDonald
University of Cambridge

Robert plans to develop and apply machine learning to help predict future flooding events in greater detail.

Climate change is bringing us more and more extreme weather, with the accompanying increased flood risk. We need to understand the extent of that risk to help us prepare the surrounding infrastructure and environment. Many hydrological models are ill-suited for predicting future macro flooding trends, particularly for data-poor regions, which are usually those most at risk from the adverse effects of climate change. Detailed predictions, however, would enable local authorities to prepare effective and lasting solutions for their communities.

Robert’s machine learning-based models will be capable of predicting flooding years from now, and at a granular level. He will train the AI to understand how variables such as rising temperatures and river flow will affect water levels far into the future. His industrial partner, Mott McDonald, a global engineering, management and development consultancy, can use the data this model will produce to advise local authorities on how to work with the elements within their control – infrastructure and water management – to help build resilience against climate change and its effects.

Robert also intends for his cooperation with Mott McDonald to produce a system that is as user-friendly as possible and incorporates user requirements from the outset.

Robert considers himself an applied mathematician, an engineer, and a designer, specialising in creative thought applied across fields. He is a Future Infrastructure and Built Environment PhD student at the University of Cambridge, working in Artificial Intelligence for Environmental Risk, as well as Chief Technology Officer at Ichthion Limited, a company he co-founded.

Robert Rouse Thumbnail
“The Industrial Fellowship programme will enable me to work faster and more flexibly with greater computational and information resources, given the huge amount of data involved,” said Robert. “It’s also enabled me to benefit from the programme’s network of inter-disciplinary experts and will bring my research closer to application.”