Effective flood preparation and response requires detail beyond river flows. We need to forecast where it will flood, and when. But the physical models used to map flood are very slow, taking up to 24 hours to run.
A pre-trained machine learning model can reduce this compute time to 1-2 minutes, forecasting hourly flood inundation maps in real-time.
Click to view Westport case study:
NIWA is now focused on operationalising this machine learning capability and extending it to other locations at high risk of flooding.
