Artificial intelligence (AI) is significantly enhancing cyclone forecasting, allowing for more accurate predictions of these severe weather events.
In Queensland, where the Bureau of Meteorology has issued a warning for a potential second tropical cyclone in just over a month, AI-driven tools are proving to be invaluable.
As of Wednesday afternoon, the bureau estimated a 60% chance that a low-pressure system in the Coral Sea could develop into a cyclone, potentially named Kirrily, by next Tuesday.
While the exact path is uncertain, Daniel Hayes, a community information officer in Cairns, indicated that if the cyclone does form, it might approach between Cairns and Rockhampton. Despite this, no direct impact on the coast is anticipated within the next week.
European Centre’s AI-based forecasts show Kirrily nearing Rockhampton, enhancing cyclone tracking accuracy.Currently, the bureau has not provided a detailed forecast track, but there is a possibility that the cyclone could intensify to a severe category three storm, with winds ranging from 118km/h to 159km/h.
The European Centre for Medium-Range Weather Forecasts (ECMWF) has used its AI-based model to predict Kirrily’s potential trajectory, showing it could be near the Queensland coast near Rockhampton by late next Thursday.
The ECMWF’s AI/Integrated Forecasting System (AIFS), launched last October, along with other AI models like Google’s GraphCast and Huawei’s Pangu, are now publicly accessible.
These models enhance traditional physics-based forecasting methods, which still provide essential data for training AI systems. However, AI models are based on single calculations, unlike traditional models that use ensemble approaches to create probability maps.
Despite advancements, Jyoteeshkumar Reddy Papari from Csiro notes that accurately predicting a cyclone’s track a week in advance remains challenging.
While AI models offer valuable guidance and are becoming more accessible, official warnings from agencies like the Bureau of Meteorology should always be followed. AI is expected to further improve forecasting and public access to weather predictions in the future.