Experts suggest that synthetic intelligence has the potential to save lives by predicting storm landfalls earlier than traditional methods.
A recent AI tool developed by Google DeepMind accurately forecasted the landfall location of Hurricane Lee in Canada three days ahead of existing predictions. The precision of temperature forecasts has significantly improved over time.
The speed and analytical capabilities of AI in processing historical data to make predictions are considered groundbreaking by experts. This advancement raises questions about the impact of changing climates on the intensity of storms.
Accurate weather predictions not only help individuals plan their daily activities but also provide crucial warnings about severe weather conditions like hurricanes, floods, and heatwaves, allowing communities to prepare adequately. However, traditional temperature forecasting methods require substantial computational resources.
The new AI tool, GraphCast, developed by Google, surpasses the performance of the world’s leading Western Medium Range Weather Forecasting model in over 90% of factors, as reported in a peer-reviewed publication by DeepMind in Science. Despite its innovative approach, GraphCast produces forecasts in less than a minute, requiring only slightly more computing power than conventional techniques.
The standard practice of weather forecasting involves collecting vast amounts of real-time environmental data from various sources worldwide. This data is then processed by sophisticated software to generate predictions on temperature changes and atmospheric conditions.
AI streamlines this process by leveraging machine learning to analyze extensive datasets, such as those from the ECMRWF model, to understand weather patterns and make accurate predictions about future changes in wind patterns.
While GraphCast’s forecasts may not be as comprehensive as those from traditional models, they excel in tracking major storms and predicting significant weather events like extreme temperatures. For example, the AI accurately predicted the path of Hurricane Lee well in advance compared to conventional forecasts.
Despite the success of AI-driven forecasts, experts like Remy Lam emphasize that these technologies should complement rather than replace traditional methods. The collaboration between AI and conventional forecasting approaches is seen as the way forward in enhancing weather prediction accuracy.
As AI continues to advance in weather forecasting, organizations worldwide are exploring the integration of AI technologies to improve prediction capabilities. The UK’s Met Office, in partnership with the Turing Institute, is at the forefront of researching AI’s potential in enhancing weather forecasts.
While AI-based tools show promising advancements in weather prediction, experts caution that their effectiveness may be limited by the challenges posed by climate change and the emergence of new weather extremes. The evolution of extreme weather events, such as the rapid intensification of Hurricane Otis, underscores the need for AI models to adapt to changing climate conditions for more accurate predictions.
