Artificial intelligence technologies could significantly reshape how national meteorological services deliver forecasts and weather-related services, according to a report from Amazon Web Services (AWS).
The report, titled How innovative technology can transform the delivery of meteorological services, examines how AI, generative AI and agentic AI could complement traditional physics-based forecasting methods such as numerical weather prediction. While NWP models remain central to forecasting, AI systems can rapidly analyze large datasets and identify patterns from historical weather data to produce more detailed and localized forecasts.
According to the report, hybrid forecasting approaches combining AI and traditional modeling are likely to become increasingly common. These methods could support improved short-term forecasts and longer-range predictions across multiple sectors.
Klemen Bergant, executive director of EUMETNET, a network of 33 European national meteorological and hydrological services, said AI is likely to influence many aspects of meteorological services: “AI will enter every segment of the meteorological value chain, from replacing some of the current forecasting instruments to developing new and more effective weather forecasting products and services.”
The report suggests that users are increasingly seeking information on weather impacts rather than forecasts alone, prompting meteorological agencies to develop services that support decision-making.
Charles Ewen, director of technology at the UK Met Office, said expectations have shifted toward actionable insights.
“These have moved beyond ‘what will the weather be’ to ‘what impact will the weather have and what actions should I consider,’” he said. “We need to learn how to work better with private-sector and tech companies to deliver composite value chains.”
The report highlights several initiatives demonstrating collaboration between public meteorological agencies and private technology companies.
One example is the WIS 2.0 data-sharing framework from the World Meteorological Organization. Launched in January 2025, the cloud-based system provides global access to real-time meteorological data and forecasting resources. It was developed by AWS in collaboration with the UK Met Office, the National Weather Service in the USA and Synoptic Data.
The UK Met Office has also been testing generative AI to modernize weather communication products such as the Shipping Forecast. Working with AWS, the organization has explored using large language models including Amazon Nova and Claude to automate drafting of forecast text products.
Another example highlighted in the report is Sencrop, a French agritech company that uses machine learning and data from weather stations to provide hyperlocal forecasting tools for farmers. The system analyzes large volumes of environmental data to support decisions on irrigation, pesticide use and harvest timing.
The report notes that wider use of AI in meteorology could support applications ranging from disaster preparedness and defense planning to airline route optimization, logistics operations and energy demand forecasting.
However, the authors also highlight governance and trust challenges, including concerns about cybersecurity risks, uneven data quality between regions and transparency in AI-generated forecasts.
According to the report, effective adoption of AI in meteorological services will depend on public-private collaboration, transparent governance and maintaining confidence in forecast data and methodologies.
Related news, Chongqing expands AI-powered weather services to improve warning times