Menu Menu
[gtranslate]

Can AI help determine when extreme weather is caused by climate change?

Heatwaves are becoming increasingly normal around the world, but not all of them are climate anomalies. Now, researchers have developed new low-cost AI system to help us understand the extent to which the climate crisis is to blame.

It’s an undeniable fact that our planet is heating up, despite London’s chilly ‘summer’ weather begging to differ.

Meteorologists declared 2023 ‘the hottest year in modern history’ and this year is on track to match or break that record, resulting in adaptation measures such as the closures of schools and new heat protection laws to be deployed around the world.

We’re finally coming to terms with our new reality, marked by high temperatures, flooding, drought, and intense storms – but how can know for certain when these events are a direct cause of the climate crisis?

Researchers at Stanford and Colorado State University wanted to help us find out. To do so, they developed a tool using Artificial Intelligence to ascertain the correlation between extreme weather and global temperatures.

Now up and running, this rapid and low-cost machine learning system can determine how the climate crisis has contributed to extreme heat in recent years. It offers greater clarity on the conditions and causes of different weather events, while their results can help guide climate preparedness and adaptation strategies.

Once it’s more widely used, it may also may help silence climate change deniers who believe that the world is simply going through another one of its natural warming and cooling cycles.

How does the AI tool work?  

Leading the study was Jared Trok, a Ph.D. student in Earth system science at the Stanford Doerr School of Sustainability.

Trok and his team trained AI models to predict daily maximum temperatures based on regional weather conditions and global mean temperatures.

To do this, they input a large database of climate model simulations spanning more than two centuries – from 1850 to 2100. Next, the researchers input the actual weather conditions from specific real-world heat waves.

This helped the AI tool to predict how hot the heat waves would have been if the exact same weather conditions occurred at different levels of global heating.

The researchers then compared these predictions at different global warming levels to understand how climate change has influenced the frequency and severity of historical weather events.

Looking at specific cases

The first job for the AI model was to analyse the 2023 Texas heat wave. This specific weather event contributed to a record number of heat-related deaths in the state last year.

The results revealed that global warming made the historic heat wave 1.18 to 1.42 degrees Celsius hotter than it would have been without the warming effects of climate change.

To test their model further, they asked the tool to predict the severity of record-setting heatwaves that took place in other parts of the world. They found that the AI tool’s results were consistent with existing studies on these other weather events.

This consistent accuracy opened up a new avenue of discovery.

Researchers asked their tool to predict how intense heat waves could become with the same weather patterns that caused previous record-breaking heat waves – this time, amplified by higher levels of global warming.

They found that the worst heatwaves seen in Europe, Russia, and India over the past half-century could happen multiple times per decade if global temperatures reach 2.0 C above pre-industrial levels.

Given that that we’re already approaching warming of 1.3° C above pre-industrial levels, this should be a major wakeup call.

Why is this kind of AI tool so important?

Jared Trok explains how his team’s tool fills a necessary gap in research related to the climate crisis.

Because the AI tool is trained with real historical weather data to make its predictions, it doesn’t require the creation of expensive new climate model simulations to provide accurate results.

That’s the beauty of Artificial Intelligence: it can be trained using a breadth of already-existing data. The use of machine learning for analysis also enables predictions and results to be developed at a much faster rate compared to human-conducted analysis.

Trok and his team are now working to develop the tool further, so that it’s able to provide information about other extreme weather events in more parts of the world.

This information will be vital to help us build effective climate adaptation strategies, while improving the predictability of when extreme climate events could occur.

It’s safe to say that when it comes to information on the climate crisis, more is definitely more.

Accessibility