UCT hydrologists are advancing flood early warning systems through innovative data-driven modelling presented at a global hydrology assembly.

Researchers from the University of Cape Town’s (UCT) Department of Environmental and Geographical Science (EGS) recently presented groundbreaking research on flood early warning systems at the XIIth Scientific Assembly of the International Association of Hydrological Sciences (IAHS 2025), held in Roorkee, India.

Dr Daniel Kibirige, a lecturer in the EGS department, and Nicholas Byaruhanga showcased the pilot testing results from their ongoing Water Research Commission (WRC)-funded project, which explores alternative methods and inputs for improving flood forecasting in South Africa.

Their presentation addressed one of hydrology’s most enduring challenges – the availability and reliability of data in gauged and ungauged catchments. To address this, the research team is integrating machine learning algorithms with the Rain-on-Grid modelling approach to enhance the spatial representation of rainfall–runoff processes and bridge critical data gaps. This innovative framework allows for more accurate and scalable simulations of flood dynamics, particularly in data-scarce regions.

“Accurate flood prediction is essential for safeguarding lives and infrastructure, especially in regions with limited hydrological data,” said Dr Kibirige. “Our work demonstrates how combining computational modelling with local insights can produce more effective and inclusive early warning systems.”

The project is being conducted in collaboration with the University of KwaZulu-Natal, the Water Research Commission (WRC), and several local municipalities, who continue to support the testing and refinement of the modelling approach.