🔗 Share this article

When Tropical Storm Melissa was churning off the coast of Haiti, weather expert Philippe Papin had confidence it would soon grow into a major tropical system.

As the primary meteorologist on duty, he predicted that in just 24 hours the weather system would intensify into a severe hurricane and start shifting towards the coast of Jamaica. Not a single expert had previously made this confident forecast for quick intensification.

However, Papin possessed a secret advantage: artificial intelligence in the form of Google’s recently introduced DeepMind cyclone prediction system – launched for the first time in June. And, as predicted, Melissa evolved into a system of remarkable power that ravaged Jamaica.

Growing Reliance on Artificial Intelligence Forecasting

Forecasters are increasingly leaning hard on Google DeepMind. On the morning of 25 October, Papin clarified in his public discussion that the AI tool was a key factor for his certainty: “Approximately 40/50 AI ensemble members show Melissa reaching a most intense hurricane. While I am unprepared to forecast that strength yet given track uncertainty, that is still plausible.

“There is a high probability that a phase of rapid intensification is expected as the storm moves slowly over exceptionally hot ocean waters which is the highest oceanic heat content in the whole Atlantic basin.”

Surpassing Conventional Systems

The AI model is the first AI model dedicated to tropical cyclones, and currently the initial to beat traditional meteorological experts at their own game. Through all 13 Atlantic storms this season, Google’s model is the best – surpassing experts on track predictions.

The hurricane eventually made landfall in Jamaica at maximum strength, one of the strongest landfalls ever documented in nearly two centuries of record-keeping across the region. The confident prediction likely gave people in Jamaica additional preparation time to prepare for the catastrophe, potentially preserving people and assets.

How Google’s System Works

Google’s model works by spotting patterns that traditional time-intensive physics-based prediction systems may miss.

“The AI performs far faster than their physics-based cousins, and the computing power is more affordable and time consuming,” said Michael Lowry, a ex meteorologist.

“This season’s events has demonstrated in quick time is that the newcomer artificial intelligence systems are on par with and, in some cases, superior than the slower traditional forecasting tools we’ve traditionally leaned on,” Lowry said.

Understanding Machine Learning

To be sure, Google DeepMind is an example of machine learning – a technique that has been employed in research fields like weather science for a long time – and is distinct from creative artificial intelligence like ChatGPT.

Machine learning processes mounds of data and extracts trends from them in a manner that its model only takes a few minutes to come up with an answer, and can do so on a desktop computer – in sharp difference to the primary systems that governments have used for decades that can take hours to run and need the largest high-performance systems in the world.

Expert Responses and Upcoming Advances

Nevertheless, the fact that Google’s model could outperform earlier gold-standard legacy models so quickly is nothing short of amazing to weather scientists who have spent their careers trying to forecast the most intense storms.

“I’m impressed,” commented James Franklin, a retired forecaster. “The data is sufficient that it’s pretty clear this is not just beginner’s luck.”

Franklin said that although Google DeepMind is outperforming all competing systems on predicting the trajectory of hurricanes globally this year, like many AI models it sometimes errs on high-end intensity forecasts wrong. It had difficulty with Hurricane Erin earlier this year, as it was similarly experiencing rapid intensification to category 5 north of the Caribbean.

In the coming offseason, Franklin said he intends to talk with Google about how it can enhance the DeepMind output more useful for forecasters by providing additional under-the-hood data they can use to assess exactly why it is coming up with its answers.

“The one thing that nags at me is that while these predictions appear really, really good, the output of the system is kind of a opaque process,” said Franklin.

Broader Industry Developments

There has never been a private, for-profit company that has developed a top-level weather model which allows researchers a view of its techniques – in contrast to most other models which are provided free to the general audience in their entirety by the authorities that designed and maintain them.

The company is not the only one in starting to use artificial intelligence to address difficult meteorological problems. The authorities are developing their own AI weather models in the works – which have demonstrated improved skill over earlier traditional systems.

Future developments in artificial intelligence predictions appear to involve startup companies taking swings at previously tough-to-solve problems such as sub-seasonal outlooks and better advance warnings of severe weather and flash flooding – and they are receiving US government funding to pursue this. A particular firm, WindBorne Systems, is also launching its own atmospheric sensors to fill the gaps in the national monitoring system.