JMA to test AI-enhanced weather forecasting
AI may help JMA identify patterns and deliver more accurate forecasts.
The Japan Meteorological Agency (JMA) is exploring the use of AI to improve the accuracy of weather forecasts, with a particular focus on deep learning technologies, according to a source familiar with the plans.
A dedicated team was launched in April to begin developing the infrastructure and tools needed to integrate AI with JMA’s existing numerical weather prediction models. The goal is to combine traditional simulations with AI-generated forecasts based on historical weather data.
If implemented, AI systems could identify weather patterns more efficiently and enhance forecasts for variables such as rainfall and temperature. The technology may also offer improved accuracy in predicting extreme weather events like typhoons.
Currently, the JMA relies on supercomputers to simulate future atmospheric conditions based on observational data. Human forecasters then review the outputs, applying expert judgment before issuing final forecasts and alerts. Even with AI integration, human oversight will remain a core part of the process.
In addition to forecasting, the agency is also considering AI for processing data from the Himawari-10 satellite, which is expected to launch in fiscal 2029.
An official announcement outlining further AI integration measures is anticipated in June.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!