Environment and Climate Change Canada is set to launch a first-of-its-kind hybrid weather forecasting model this spring that combines artificial intelligence with traditional physics-based methods, the federal department announced.
The move positions Canada as what Jean-François Caron, a research scientist with Environment Canada, described as the first meteorological service in the world with a public forecast partially generated using AI.
The hybrid system pairs AI’s ability to detect large-scale weather patterns with the precision of traditional physics-based modelling. According to the department, AI can analyze decades of historical data covering an entire continent in minutes, identifying relationships between temperature, wind, and pressure to estimate future atmospheric conditions.
However, AI modelling alone has its limits.
“However, they lack the fine-scale details, like high wind gusts, the high temperature, or very local precipitation,” Caron told PortageOnline.com. “(AI) struggles to accurately predict those.”
To address this, the physics-based component – Environment Canada’s Global Environmental Multiscale model – handles the smaller-scale details that AI tends to miss, such as local wind gusts and precipitation.
Caron also sought to distinguish this application from mainstream AI tools.
“When you say AI, people think about ChatGPT,” he said. “Here, it’s something completely different.”
Rather than generating text, the AI produces data that human meteorologists then interpret to issue public forecasts.
The practical gains for Canadians are significant. According to the department’s news release, the hybrid model is expected to deliver predictions of major systems, such as winter storms, heat waves, and atmospheric rivers, from eight hours to more than 24 hours earlier than current models allow.
It also said a six-day forecast under the new model will be as accurate as a five-day forecast under the traditional system.
“It’s really from day three to day seven for the public; that there will be big gains in terms of accuracy,” Caron said. “We’re saying that a six-day forecast with the new model will be as good as a five-day forecast from the previous model for us; it’s a huge change, and we are very proud.”
Despite the integration of AI, the department said meteorologists are not being replaced. Caron noted the shift will not eliminate jobs and may, in fact, create positions on the research side.
Not everyone, however, is fully convinced of the model’s potential. Halifax-based meteorologist Cindy Day, who has more than 40 years of experience, raised concerns about the reliability of historical data in an era of rapid climate change.
“The rate at which our temperatures are changing and our climate is changing is significant,” Day told CBC News. “And so I’m not sure that the analysis of so much data going back so much time so very quickly is going to make a big difference into producing a forecast for the next five, six, seven days.”
The department said the official launch is expected in the second half of May 2026.
