Around 604 extreme weather events were recorded worldwide in 2024, and many were described as unusual or unprecedented. The EU-funded XAIDA project set out to show how climate change shapes such events and how they can be better understood. Using machine learning and deep learning, the researchers found a worrying result: in many regions, heatwaves are increasing faster than climate models expected – and at the end, you can check how well you handle the heat yourself.
Extreme weather events have become more frequent and more intense in recent years. That raises serious concerns, both for the environment and for human safety.
The numbers speak for themselves: in 2024, around 604 such events were recorded worldwide, and many of them were classified as unusual or even unprecedented.
The link between these events and climate change looks strong. The trouble is that the mechanisms behind their growing intensity and complexity are still not fully understood. This is the gap that the XAIDA project set out to close.
XAIDA is an EU-funded project that brought together 16 partner organisations specialising in three fields: artificial intelligence, statistics and climate modelling.
Their goal was to develop data-driven methods that reveal the role climate change plays in extreme weather around the world. Stakeholders from different sectors were also involved, so that the findings could feed into risk assessment and adaptation strategies.
The consortium developed a set of tools for machine learning, causal inference, event detection and impact assessment. Heatwaves were one of the key areas where these tools were put to work.
The team used variational autoencoders, a deep learning technique, to examine how extreme heat behaves as the climate changes. The results were not reassuring.
“During the project, we learned that in many regions, heatwaves have been increasing faster than anticipated by climate models. This is a major concern,” says Dim Coumou, the project coordinator, based at Vrije Universiteit Amsterdam.
XAIDA leaves behind several concrete results:
The project also produced more than 100 peer-reviewed scientific papers and developed educational materials for primary and secondary school teachers.
One of the biggest challenges concerned AI itself. Its results need to be trustworthy and interpretable, otherwise scientists have little reason to rely on them.
XAIDA tackled this by combining traditional methods, such as climate models and statistical analysis, with new AI techniques. Running the two side by side made it possible to check results against each other and validate them.
This mix of the old and the new is what allowed the team to present AI-based findings with confidence rather than as a black box.
Coumou stresses that XAIDA did not happen in isolation: “XAIDA has been part of a much larger wave of AI innovation in climate science in recent years; AI has really taken off during the lifetime of the project. Now we have global AI weather prediction models that are really providing the research community with new opportunities.”
For anyone thinking about a career in climate research, the message is clear: studying the weather today also means working with data, statistics and algorithms.
XAIDA brought together artificial intelligence, statistics and climate modelling to better understand extreme weather. The project found that heatwaves in many regions are rising faster than climate models predicted, and it delivered the AIDE toolbox, the CAUSEME platform and GreenEarthNet. By pairing traditional methods with AI, the team was able to validate its results and make AI more trustworthy for climate science.
5 questions · one minute
EN 
CS
DE
ES
FR
HU
IT
PL
PT
РУ
SK
TR
УК
AR
中文