Flash flooding remains one of Australia’s most dangerous and difficult to predict natural hazards. Developing within hours of intense rainfall, these events can pose significant risks to public safety and challenge emergency response efforts.
Natural Hazards Research Australia (the Centre) and the Bureau of Meteorology (the Bureau) researchers are using case studies and detailed analysis of high-impact rainfall events to strengthen Australia’s future forecasting capability and support more effective emergency management.
The May 2018 flash flood in Hobart provides an important case study for improving how extreme rainfall and inundation are forecast with insights being shared at the Australian Meteorological and Oceanographic Society (AMOS) Conference in Adelaide, 16-20 February.
Dr Paul Fox-Hughes from the Bureau and the Centre will share insights from the Developing an integrated predictive capability for extreme rainfall and inundation project. He said the project provides the opportunity to draw together several threads of Bureau research into better forecasting heavy and intense rainfall and communicating the associated uncertainties.
'This is always important but particularly so now, when we have seen the consequences of rapid onset heavy rainfall events play out across the breadth of Australia in recent weeks,’ said Paul.
Improving prediction of dangerous rainfall
Flash flooding is inherently difficult to forecast because of uncertainties in where, when and how intensely rainfall will occur, and how landscapes respond. To address this challenge, the project combines probabilistic weather modelling, short-term radar-based forecasting and emerging machine-learning techniques to better predict the likelihood and severity of flash flooding.
Testing these approaches against real events, including the 2018 Hobart flood, allows researchers to evaluate how accurately new high-resolution weather models simulate extreme rainfall. Preliminary findings show strong agreement between modelled rainfall and radar observations, demonstrating promising capability for predicting high-impact downpours.
Together, these advances are helping lay the groundwork for next-generation operational forecasting systems and more reliable early warnings for communities and emergency services.
Beyond improving meteorological science, the project is designed to ensure research outcomes are practical and usable.
Close engagement with the Bureau and emergency agencies shapes the tools that will better communicate uncertainty, provide probabilistic guidance, and support real-time operational decision making during severe weather.