Now,

the idea that maybe we can do predictions, some kind of prediction of the weather and the atmosphere just based on data has been around for a long time, even before World War II, before having these supercomputers.

If people are watching this movie called Pressure, I mean, it sort of goes to some of the history of when people were just looking at some weather maps and making weather predictions.

So whether that can be used for weather prediction and even climate prediction, that idea has been around.

It kind of went away when physics-based modeling showed so much success in the 50s and 60s until recently.

But now that weather forecasting with AI has been so successful, just starting from 2022, people are looking at whether some of that kind of technology can be used for climate prediction.

What are these AI weather models? So they use the same technology that we have in chat GPT and these kind of large language models, but at the end it's different.

It would fill a very large gap currently in our ability to fully span the spectrum of mitigation and what the implications are for different pathways that we might choose.

Now,

the idea that maybe we can do predictions, some kind of prediction of the weather and the atmosphere just based on data has been around for a long time, even before World War II, before having these supercomputers.

If people are watching this movie called Pressure, I mean, it sort of goes to some of the history of when people were just looking at some weather maps and making weather predictions.

So whether that can be used for weather prediction and even climate prediction, that idea has been around.

It kind of went away when physics-based modeling showed so much success in the 50s and 60s until recently.

But now that weather forecasting with AI has been so successful, just starting from 2022, people are looking at whether some of that kind of technology can be used for climate prediction.

What are these AI weather models? So they use the same technology that we have in chat GPT and these kind of large language models, but at the end it's different.

It would fill a very large gap currently in our ability to fully span the spectrum of mitigation and what the implications are for different pathways that we might choose.

Now,

the idea that maybe we can do predictions, some kind of prediction of the weather and the atmosphere just based on data has been around for a long time, even before World War II, before having these supercomputers.

If people are watching this movie called Pressure, I mean, it sort of goes to some of the history of when people were just looking at some weather maps and making weather predictions.

So whether that can be used for weather prediction and even climate prediction, that idea has been around.

It kind of went away when physics-based modeling showed so much success in the 50s and 60s until recently.

But now that weather forecasting with AI has been so successful, just starting from 2022, people are looking at whether some of that kind of technology can be used for climate prediction.

What are these AI weather models? So they use the same technology that we have in chat GPT and these kind of large language models, but at the end it's different.

It would fill a very large gap currently in our ability to fully span the spectrum of mitigation and what the implications are for different pathways that we might choose.

It does follow certain patterns.

Now, when AI started really to change many, many fields in medicine, in science, and in engineering, some people in my field in around 2017 started thinking, OK, what can you do for climate and weather? And so my group was one of the first groups who started thinking about this question.

Now, well, weather has patterns, right? And these methods are amazing in pattern recognition.

So how about we just try to...

Go back to this idea that weather can be predicted as an image processing problem.

And I should maybe take a step back and say that before 1940s, weather forecasting was a pattern matching problem.

This is actually how weather forecasting was done during World War II, for example, for the Normandy landing, that people had this catalog of weather patterns.

I want to know where Pedram sees this making the biggest difference.

Pedram Hassanzadeh, a geophysicist at the University of Chicago and one of the article's authors, explains what gray swans are and why they may be challenging for AI to

predict.

AI weather models are really transforming weather forecasting.

They learn from the past patterns.

So they are doing a very good job predicting things that are not that different from what they have seen before.

The question is, if these models haven't seen something in the past, can they still predict them in short-term forecasting? Grace 1 are things that are physically possible.

We just haven't seen them.

If we only add data back in the West Pacific region, the model can learn from that and predict strong hurricanes in the West Atlantic Ocean.