Can you run the Olympics 10,000 metres final in 50-degree heat? With once-a-century weather now taking place once a decade, the Paris 2024 Olympics organisers are wrestling with unprecedented climate scenarios for the world’s oldest global sporting contest. But with climate change making weather forecasts harder to predict, they’re relying on new climate modelling techniques for more accurate predictions.
In a world experiencing climate change, we want to receive warnings before unprecedented events happen. We therefore account for climate change in both our long-term climate models, which help guide how we build our cities and design our infrastructure, and also our short-term weather forecasts.
“In a world experiencing climate change, we want to receive warnings before unprecedented events happen,” says Erin Coughlan de Perez, Dignitas Professor at Tufts who specialises in modelling climate extremes. “We therefore account for climate change in both our long-term climate models, which help guide how we build our cities and design our infrastructure, and also our short-term weather forecasts, which provide a heads-up before extreme events happen, so we can prepare and avoid as many impacts as possible.”
But there’s a Catch-22. Although the rise of supercomputers and higher-resolution climate models has improved weather forecasters’ ability to predict an extreme weather event, it can be difficult for forecasters to believe their own extreme weather predictions due to their extremity. It’s like trusting the anomalous result on a graph.
“In recent years, extreme weather events have come as a surprise to both the public and experts by smashing previous records by large margins,” says Dr Erich Fischer, Professor for Weather and Climate Extremes at ETH Zurich. Due to the increasing regularity of extreme weather events, such as the Pacific Northwest heatwave and heavy rainfall events in North-western Germany in 2021, new climate model techniques are needed that can simulate unprecedented scenarios. This way, “stakeholders and politicians will be able to adapt and prepare before an event, not in its aftermath,” says Fischer.
How do you predict the weather with no historical precedent? One new approach, known as UNSEEN (Unprecedented Simulation of Extremes with Ensembles), has its roots in the 2014 flooding in the UK, which over a century of British records didn’t predict. The UNSEEN method re-ran the models – using historical data from previous British winters – but this time searching for extremes at various points to see what would happen later down the line.
“Thinking about the most extreme heatwave or flooding to occur in the next few years is scientifically challenging to answer,” says Fischer, who created a method similar to UNSEEN called ‘ensemble boosting’, which he likens to the “butterfly effect”. “Ensemble boosting is based on the same climate models as traditional methods, but we push the model to the limit enabling us to sample the most extreme cases possible. We run the model for a long time, look at the worst case and make it even worse. We then start the model again and see where they end up 2-3 weeks before the event like the Olympics and whether the prediction is physically consistent and reasonable.”
But prediction is not the only aim of new modelling approaches with some researchers also looking to prove global warming’s role in producing today’s extreme weather events. For example, one method uses sea level air pressure records dating from global warming's infancy in the 1950s to simulate whether a recent extreme event could have been possible or likely in the pre global warming era.
I still can’t imagine 50 degrees in Paris in the next few years. That hasn’t happened anywhere in Europe yet – Sicily came close, but Paris is a lot further north. However, it could easily be 2-3 degrees more than it's been and if that happens over 10-15 days, that would be a huge stress on cooling and energy systems.
While inaccurate weather predictions are annoying for most of us, it’s not life or death, unlike organisers of worldwide sports events tasked with keeping 15,000+ athletes and 15 million spectators cool in the case of an unprecedented heatwave. In particular, organisers need accurate predictions to create a resilient ‘heat plan’ and stress test the city’s cooling and energy systems.
“The City of Paris approached scientists and asked whether 50 degrees would be possible,” says Fischer. “They wanted to know how hot it could get and how persistent that heatwave could be. As it’s very challenging to predict, we have combined different approaches, such as ‘ensemble boosting’ and other statistical models, including sampling from old UK Met Office weather forecasts of worst cases that never happened.”
The Paris Olympics is particularly vulnerable to a heatwave as it’s taking place during that city’s hottest weeks, with a recent study finding temperatures could be up to four degrees higher than the 2003 European heatwave. So, what does Fischer think?
“I still can’t imagine 50 degrees in Paris in the next few years. That hasn’t happened anywhere in Europe yet – Sicily came close, but Paris is a lot further north. However, it could easily be 2-3 degrees more than it's been and if that happens over 10-15 days, that would be a huge stress on cooling and energy systems.”
While predicting unpredictability is inherently unpredictable, we hope that the athletes who have prepared for years and timed their peak fitness for this event don't feel like they're underperforming due to extreme heat.
