The Future of Weather Forecasting | Breakthrough

JOE SIENKIEWICZ: So I started out 28 years ago. Just imagine, forecast information came in the form of paper, piles of paper. It limited the amount of information that we could look at. We see things now in the models that we're actually, in some ways, learning, and confirming using other information, observation satellite data. So things have changed enormously in my career.

NARRATOR: Today forecasters can rely on a vast array of weather sensors on the ground, at sea, in the air, and even space. This information is fed into computer models that build on our deep knowledge of environmental physics.

WILLIAM LAPENTA: The atmosphere is basically a fluid, just like water, fish tank. And fluids are defined by mathematical equations in terms of their structure and how they would evolve with time. So we go for the major terms in an equation, and then we put them into a computer model. And we try to get a solution of how the atmosphere will evolve.

NARRATOR: Dr. William Lapenta is the director of NOAA's nine weather prediction centers. When a tornado is forming or a hurricane is brewing, the accuracy of his models can be a matter of life and death. But because there are so many variables in the atmosphere, no model spits out the right prediction every time. NOAA compensates by running their models dozens of times, introducing random variations in the data. The result is a cluster of possible futures called an ensemble.

WILLIAM LAPENTA: So let's just say that my putt was a hurricane track. OK? So I put once. You see the ball roll and you see it roll a certain way. OK. That's one piece of information. So an ensemble means you would do that many times over, maybe 20 times, maybe 40 times. And the thing is, every time you put it down and stroke it, something changes. So then if the trajectories of the balls are very close to each other, that means you either have a very good model, or there's a high level of predictability in that hurricane track, or both.

These are called spaghetti plots. And it's essentially just deterministic guidance overlaid with a bunch of ensemble guidance. So you can really see a lot of information on one screen. When they're closer together, you know you have higher confidence in your forecast. There's more certainty in your forecast. And then as you go out in time, solutions begin to diverge. We become more uncertain about the forecast. And really just the slightest difference can make a huge change in the forecast.