DeepMind and UK’s Met Office use AI to improve weather forecasts



Artificial intelligence improves the accuracy of short-term weather forecasts and in particular the prediction of storms and heavy rain, according to research conducted by the British Meteorological Office and DeepMind, the London-based AI company.


Their project focused on “new broadcast” – locating the time, location, and intensity of high-resolution precipitation up to two hours ahead – which is not well handled by the supercomputer models used to predict larger-scale weather over the next day. or week.

This is crucial for applications ranging from the warning of imminent flood risk to letting organizers of outdoor events know when a downpour comes.

The results, published in Nature, show that an AI-based approach called “deep generative modeling” or DGM surpassed other nowcasting methods over a wide range of measures. It was ranked first for accuracy and utility by 89 percent of a panel of 56 professional meteorologists dazzled by the source of the predictions.

“Improving the accuracy of short-term forecasts is an incredibly important effort,” said Niall Robinson, Head of Partnerships and Product Innovation at Met Office, the UK’s National Weather and Climate Service.

“Extreme weather has catastrophic consequences, including loss of life, and as the effects of climate change suggest, these types of events will become more common,” he added. “This research shows that the potential AI can offer as a powerful tool to improve our short-term forecasts and our understanding of how our weather patterns evolve.”

DeepMind and Met Office researchers trained DGM to predict the evolution of precipitation (rain and snow) by analyzing three-year British radar maps. These show how much rain falls every five minutes at 1 km spatial resolution.

AI analyzes the last 20 minutes of radar observations and predicts precipitation for the next 90 minutes © DeepMind

According to Shakir Mohamed, a senior DeepMind researcher, the AI ​​technique involved is quite different from two other areas where DeepMind has been successful: playing games like Go and determining the shape of protein molecules. The company was founded in 2010 and acquired by Google in 2014.

DGM focuses on the likelihood of sequences as precipitation patterns unfolding rather than achieving a specific outcome, such as winning a game or discovering how a protein unfolds.

Robinson said the Met Office was considering how to use DeepMind research in its operational forecasts. “We need to carefully consider how new tools are implemented and maintained, the best user interfaces for our meteorologists, and how it fits into all the other forecasts we provide,” he said.

Going forward, Robinson said the Met Office “is currently investigating where we can collaborate next on other research issues. [with DeepMind], including using AI to assess the impact of climate change and the treatment of weather observations to create even better weather forecasts. ”

The researchers said the goal was not to replace human meteorologists with automated experts but to improve their work.

“AI can be a powerful tool that enables forecasters to spend less time trawling through ever-growing piles of prediction data and instead focus on a better understanding of the implications of their forecasts,” Mohamed said. “This will be an integral part of mitigating the negative effects of climate change today, supporting adaptation to changing weather patterns and potentially saving lives.”

Source: https://www.ft.com/content/602235aa-7039-472a-80cf-55fa3519ea06

No comments

intech company. Powered by Blogger.