Artificial intelligence models have provided an additional “tool in the toolbox” for meteorologists working to predict hurricanes in the Atlantic.

The US-based National Hurricane Centre began to incorporate AI models into its forecasts last year.

Wallace Hogsett, a science operation officer with the NHC, said in a post on the organisation’s website this month that the use of AI in storm forecasting was accelerating, with new tools emerging quickly.

He said that the NHC had partnered with Google DeepMind to develop a new AI hurricane forecast model that was used “experimentally” in 2025.

The National Oceanic and Atmospheric Administration’s Environmental Modelling Centre and the European Centre for Medium-Range Weather Forecasts have developed AI-driven global forecast systems that were “available for evaluation” last year.

Mr Hogsett said: “More AI-based tools and models are coming, and we’re evaluating them all thoroughly for potential integration into forecast operations.

“At NHC, we are just beginning to scratch the surface of how these new models may be used.”

He explained that traditional numerical weather prediction models use atmospheric conditions to predict future conditions, while AI models use relationships learnt from history to deliver a forecast.

Mr Hogsett said: “Despite these differences, traditional NWP models and AI models are highly complementary.

“Because they use different methodologies, their forecasts have different types of errors, which can be very valuable to forecasters.

“These differences can be very helpful for forecasters to understand the full range of uncertainty and communicate the risk more effectively.”

He said that in 2025, AI models quickly homed in on the likely track and strength of Hurricane Melissa, which he described as difficult to forecast.

Mr Hogsett said: “It is just one example, and there are other examples where the traditional models performed better.

“However, it was promising to see the new tools performing well. This is all part of the learning process for forecasters, and we’re still in the early stages.”

He added that AI would not replace human forecasters, stating: “None of the models are perfect, and they never will be.

“Now more than ever, we need trusted experts in the loop to observe, synthesise and make sense of the vast amounts of information.

“AI will likely assist in efficiently synthesising information but without experts in the loop to constantly evaluate and integrate new tools and technologies and communicate a coherent risk-based message, lives and property would be at greater risk.”

Francis Kredensor, the Bermuda Weather Service deputy director, said it was important to note that the AI techniques were just another “tool in the toolbox” for meteorologists.

He added: “It is also imperative to note that many of these AI tools and models are quite new, so the meteorological community is still assessing their forecast accuracy against what actually happens.

“The AI versions of the NOAA Global Forecast System model and the European Centre’s ECMWF model were first available for last season, so they are still very much in an ‘evaluation’ phase.”

Mr Kredensor said that it was too soon to say if the new models would have an impact on forecast accuracy for Bermuda.

“We would want several seasons’ worth of data to see if there is a clear trend towards more accurate track and intensity forecasts, but we are excited to see if these new tools can help,” he explained.

“The same would be true of general day-to-day weather forecasting, as well — there is a lot of promise for improvement, but we are still very early in this new era of AI-driven weather models.

“Given that these tools are so new, the BWS hasn’t used them much to date, but as the AI models improve and the data becomes increasingly available, it’s something we’ll be keeping an eye on to use in the future.”