Experts from the Met Office and the Turing Institute discuss the role of AI in weather forecasting, shining a light on the work of Google, IBM and others.
Weather prediction, traditionally a vastly complex task made possible by using data from satellites and surface sensors, has seen notable advancements over the past few decades.
And now, as we take our first steps into the AI era, weather forecasting may become even easier and more accurate than ever before.
Companies like Google, IBM, NASA and Previsico are all utilising AI’s data processing capabilities to make improvements to meteorology right now.
These new forms of climate tech are touted to play a role in humanity’s response to climate change, so their development is crucial.
That was the message at a recent London conference, hosted by the UK’s Met Office, at which the Met Office and the Alan Turing Institute unveiled their own FastNet AI forecaster.
There, Professor Kirstine Dale, Chief AI Officer at the Met Office said: “We’re in the midst of an AI revolution and it’s happening at just the right time.
“There are some major environmental challenges facing us and we’re increasingly aware of our vulnerability to extremes in the weather.”
But how exactly are companies and organisations integrating AI into meteorology, and what will forecasting look like in the near future?
How are companies currently using AI in meteorology?
Two of the biggest organisations using AI in these ways are Google and IBM, both of which are leveraging AI to advance weather forecasting and climate analysis.
Let’s start with Google, which has expanded its AI-driven flood forecasting system to cover over 100 countries, potentially benefiting 700 million people worldwide.
The system offers improved accuracy with a seven-day lead time and incorporates more labelled data and a new model architecture.
Google is also making its flood forecasting data available through an API and dataset to facilitate research and response efforts, which could be crucial for future disaster responses.
IBM, in collaboration with NASA and Oak Ridge National Laboratory, has developed an AI foundation model for weather and climate forecasting.
This versatile model creates targeted forecasts, detects severe weather patterns, improves spatial resolution in climate simulations and has the capacity to enhance the representation of physical processes in weather models.
Most importantly of all, both companies are making their technologies accessible to researchers and partners, which typifies the collaborative spirit necessary for joined-up climate action.
“Our mission is to use AI to make flood forecasting information globally accessible,” says Yossi Matias, Vice President & Head of Google Research.
The positives and negatives of using AI in forecasting
The integration of AI with existing physics-based numerical models presents both opportunities and challenges.
While traditional models require significant computational resources, AI can process large datasets rapidly, offering a much more efficient alternative.
Dr Scott Hosking, Interim Director of Science and Innovation of Environment and Sustainability at the Turing Institute, notes the progress that’s already been made.
“In just a few months, the partnership between the Met Office and Turing has built something that matches the performance of traditional models,” he says.
“We are really pleased with our progress but there’s a lot more to do.”
But despite these technological advances, experts agree that the role of human expertise is indispensable to meteorology.
Dan Travers, Co-Founder of Open Climate Fix, is a big proponent of this notion. “I’m a massive believer that the computer should follow the subject matter experts into the field,” he says.
“The job of the meteorologist might change but if you just try to solve a problem with a computer alone, you’re going to miss a huge amount of insight.”
Dr Florence Rabier of the European Centre for Medium-Range Weather Forecasts thinks along similar lines as Dan. “Only meteorologists can truly assess the quality of the models to help us improve them,” she says.
The importance of collaboration
In a similar way to the models used by Google and IBM, the partnership between the Met Office and the Alan Turing Institute is all about leveraging AI for public benefit.
This particular initiative will focus predominantly on the UK, but over the next few years it is likely that AI prediction models will provide worldwide coverage, thanks to the innovative work of scientists and the investment of companies.
Dr Jean Innes, CEO of the Alan Turing Institute, says: “The aim is to put an AI weather prediction model in the hands of Met Office forecasters within 12 months.”
Likewise, Professor Penny Endersby, CEO of the Met Office, speaks about the critical importance of continuing to improve weather and climate understanding to “keep people safe, protect businesses and improve our health”.
