AI improves understanding of links between climate change and extreme weather

Fifteen European research institutes are joining forces to investigate the links between climate change and extreme weather. New methods in the field of artificial intelligence will help the climate scientists to achieve their goal. In September, the team led by Robert Vautard of CNRS-IPSL in Paris and Dim Coumou of VU Amsterdam will launch the European XAIDA project.

06/07/2021 | 12:05 PM

We have experienced more and more extreme weather in the past decade. Regions all over the world have been exposed to periods of extreme heat, drought, heavy rainfall, severe forest fires, hurricanes, etc., which cause major damage and many fatalities. What role does climate change play in these extremes?

Extreme weather can provide insight into future climate conditions and their possible effects, but not all extremes say something about the future. To be able to make such predictions, a causal link between an extreme and the human influence on climate change must first be demonstrated.

New methods needed
While climate scientists have successfully demonstrated causal links between climate change and extremes such as heat waves and heavy rainfall, it remains a challenge to prove the existence of such links in other extreme weather events.
Recently developed methods in the field of artificial intelligence offer great opportunities to fill this gap.

“AI is helping us to classify extreme weather such as tropical storms and identify the causes behind the extremes,” explains Dim Coumou. Various research sectors will need to join forces and develop new methodologies to achieve the common goal of reducing the risk of extreme weather.

Using machine learning
This is the goal of XAIDA, an EU-funded project (Horizon 2020) involving a consortium of 15 universities and research institutes and led by CNRS-IPSL and VU Amsterdam. The consortium brings together experts of machine learning, statistics and climate modelling. Together, they will design and apply new methods to study recent weather extremes, identify links with climate change, and try to predict whether or not such events will become even more extreme in the future.

The researchers will engage various stakeholders to reduce the risks and they will develop material for secondary schools to ensure younger generations are better informed. Dim Coumou will focus on making new AI methods applicable to extreme weather research and he will study the recent summer droughts in Europe.