The pros and cons of using AI to predict weather

The conversation around Artificial Intelligence has increased over the last few years, and for good reason; when it comes to weather, AI shows a lot of promise.

Normally, when we think of AI, we think of altered pictures and videos, but technology is making it possible to expand our ability to accurately predict the weather.

Traditionally, meteorologists use computer models to build a forecast called “Numerical Weather Prediction.”

These models use data from atmospheric observations, like the current air temperature and wind speed, which then get plugged into complex equations to compute future weather conditions.

The same data is used in AI-based weather models, but months and years worth.

With that data, it creates a forecast based on statistics, but like all things, even AI has its pros and cons.

---

Pros of AI:

• Models require less computational power
• Models are faster and therefore can expedite nowcasting A.K.A forecasting the immediate future
• Generate useful simulations of extreme weather

Cons of AI:

• Unintended bias in forecasts depending on data
• Less likely to predict rare events
• Doesn’t include probability
• AI machines can only analyze the data they are provided

---

There’s still development to be had before AI is used for the daily forecast, but it is already being used and developed within several weather forecasting and communication efforts.

The European Model, for example, used often in tracking developing hurricanes and tropical systems, has recently received an update.

The AI European Model, new this year, is outperforming its traditional model in several areas.

The National Hurricane Center’s Jamie Rhome considers AI to be a component of the agency’s track forecasting technique.

“We’re using it to help the forecasters sift through this rapid expansion of data because there’s so much information now that they don’t have time to get through it all. The AI kind of helps them highlight where to focus,” said Rhome.

This technology is still in its early stages, but it has already begun to prove itself as a useful tool and shows promise for the future of weather prediction.