Leveraging AI for Precision in Hurricane Forecasting 

Since 1980, the United States has faced 363 billion-dollar weather disasters. Hurricanes have caused the most damage, with total losses exceeding $1.3 trillion. On average, every hurricane event racks up a staggering $22.8 billion in losses. The human toll is just as devastating. Hurricanes have claimed 6,890 lives in the U.S. between 1980 and 2023. Globally, the impact is even more tragic, with both financial losses and death tolls soaring higher. 

As climate change fuels more frequent and intense storms, coastal communities face growing challenges. In some cases, hurricanes have reached far inland. Accurate tools are essential to predict the severity, timing, and location of these storms and assess the damage they leave behind.

We may already have the technology to make this happen. When Hurricane Beryl was rushing across the Atlantic in July, Google Deepmind’s GraphCast predicted that the storm would take a sharp turn away from southern Mexico to Texas. This forecast was made a week earlier than traditional methods, and it turned out to be accurate.

Several other major tech companies have launched AI-powered weather forecast tools. NVIDIA, for example, developed StormCast, which was developed in collaboration with the Lawrence Berkeley National Laboratory and the University of Washington.

Microsoft introduced Aurora Atmosphere, a powerful weather prediction platform that leverages 3.3 billion parameters for highly accurate forecasts. 

Building on similar advancements, NASA and IBM have collaborated to launch the Prithvi Weather and Climate AI model, which uses AI to fill gaps in climate data and improve hurricane forecasting. This model has been kept open source to allow researchers and developers to contribute and adapt the model to various applications, including hurricane forecasting. 

One of the most challenging aspects of predicting hurricanes is to understand why some hurricanes rapidly intensify. Michael Scott Fischer, a researcher at the University of Miami Rosenstiel School of Marine, Atmospheric, and Earth Science, is employing AI to gain a deeper understanding of the complex factors that drive this rapid intensification. 

To improve the accuracy of predictions, Fischer is improving current storm data by incorporating historical hurricane measurements. This includes a variety of data, such as wind speeds and humidity levels, collected from sources like buoys, satellites, and hurricane hunter aircraft. 

“I use a variety of observational data sets and data science methods to identify commonalities among subsets of storms that have undergone rapid intensification,” explained Fischer. “And hopefully, with that knowledge, we can provide tools that would improve our prediction of tropical cyclones.” 

According to Fischer, analyzing decades of storm data manually would be highly time-consuming, and this is why he is using machine learning (ML) algorithms. One of his goals is to build a 3D structure of an active hurricane in the Atlantic basin without having aircraft fly into the storm system. 

“Airborne reconnaissance over the Atlantic is routine once a storm is located west enough that we can get within range of aircraft. But even then, we don’t always have planes out there. The hurricane hunters can spend hours flying within a storm, taking readings. But eventually, they must return to base, refuel, and deploy a new crew. And that takes time,” Fischer said.

“Our method would allow us to always have a full three-dimensional structure of what a storm looks like. And it would be particularly helpful not only in the Atlantic basin but also in other basins worldwide where hurricanes form but where there are not always routine aircraft observations.” 

AI models are making significant strides in weather forecasting, but they still face notable challenges. A key limitation is their reliance on historical data. If the data used to train these models is incomplete, outdated, or inaccurate, the forecasts can be skewed. 

While AI models process data much faster than manual methods, they are still dependent on the quality of the data they receive.  Data collected from radars and planes often have noise which can undermine its quality. Additionally, the AI models may struggle to account for rare or unprecedented events that haven't occurred before.

Fisher emphasizes that despite the rapid advancements in AI technologies, it will never replace humans. Forecasts must be translated into actionable guidance, such as evacuation plans and safety precautions, to ensure communities understand the potential impact and respond appropriately. This human touch is vital for saving lives and protecting communities during severe weather.