Hey there, I'm Flora Lickman and you are listening to Science Friday.

Today in the podcast, a revolution in weather forecasting and one of the scientists behind it.

We have much more confidence in a prediction of climate 100 years from now than we have in 100 days weather.

No. climate 100 years from now, then we have in 100 days weather. Nowadays, the weather is right at our fingertips.

It's definitely one of my most visited apps.

But until about 40 years ago, scientists couldn't forecast the weather beyond 10 days.

And that means they couldn't predict seasonal patterns.

So you couldn't know how hot it was going to get in the summer or how much snow to expect in the winter.

And without that, you can't plan for things like wildfires, droughts, or hurricanes.

But climate scientist, Jagadishukla, who grew up in a village in rural India and saw how people's lives depended on this information, made it his mission to forecast seasonal weather events.

And in doing so, he changed the course of modern weather prediction.

He tells the story in his new book, A Billion Butterflies, A Life in Climate and Chaos Theory.

He's a distinguished professor at George Mason University in Fairfax, Virginia,

and he was one of the lead authors on the 2007 report of the Intergovernmental Panel on Climate Change,

which was awarded a Nobel Peace Prize.

Shukla, welcome to Science Friday.

Thank you.

Let's start in the present. How good are we at

predicting the weather today? I think that our three-day, five-day forecasts are really pretty good.

30 years ago, we will not put much stock in the actual forecast five days ahead. So thanks to this revolution

in computing, in getting better models, getting a lot of observations from satellites,

I think the weather prediction has been steadily improving. And up to seven days, it's pretty good. But then it gets a little bit tricky

...