White: How artificial intelligence is changing weather forecasting | News

White: How artificial intelligence is changing weather forecasting |  News

The University of Albany professor of atmospheric and environmental sciences studies extreme weather and weather analysis and forecasting. In a project last summer, the University of Albany researcher, who has worked for 16 years and a team of faculty and students, sent up more than a dozen balloons amid severe weather events. Their goal was to collect data on how environmental factors in upstate New York’s diverse landscape contribute to major storms, such as the increasingly common floods that have caused severe flooding in our region.

“We have to get above the roof,” 40-year-old Lazear told me. “Releasing a weather balloon opens up some mysteries about what happens just above the surface.”

Lazear said the data collected is partly helping to change the way forecasters can do their work. But weather balloons actually only scratch the surface of what happens inside meteorology.

Last week, Lazear, who lives in Albany, traveled to Baltimore to attend the American Meteorological Society’s annual conference, an event attended by nearly 7,000 scientists whose interests range from solar flares to space weather to cloud physics.

As NPR reported about the conference, meteorologists face increasing pressure to improve their forecasts now that severe weather events are becoming more common.

Lazear emerged from the conference convinced that the outlook was improving dramatically—and quickly.

Great motivator? artificial intelligence.

Artificial intelligence has already been integrated into the work of the University at Albany’s Department of Atmospheric Sciences, and that is expected to increase in the coming years, Lazear said.

Here’s part of our conversation, lightly edited for clarity, as Lazear takes me below the surface regarding the ways in which AI is expected to change meteorology forever.

White: What was your biggest takeaway from the conference?

Lazear: To be sure, artificial intelligence is rapidly changing atmospheric science. Historically, when we make weather forecasts and look at weather models, for example, we’re looking at a model built on an understanding of the fundamental physics of the atmosphere. How does temperature change over time? How does he respond to stress? And you run a super-complex system of equations on a supercomputer to calculate all of that and predict the weather, which is an amazing and complex problem. With AI, instead, you can give it a training dataset, or a feed of all observed weather conditions, to make predictions.

Witt: What are the implications of that?

Lazear: There is still a long way to go, but it is moving so quickly that I believe that in just a few years, the world of weather modeling will look very different.

White: This means increased accuracy?

Lazear: Within a few years, we will see an increase in accuracy, especially in long-term forecasts, forecasts for 5 to 10 days and beyond.

White: This is becoming more important, as extreme weather becomes more common, as your research shows, right?

Lazear: exactly. I think we’ve done a good job in the short-term outlook. Like if you want to forecast for the next 24 to 48 hours, I think we’re doing a pretty good job, and machine learning will probably help in that range as well. But I think what we’ll really see a big improvement in is medium to long-term forecasting, sometimes referred to as seasonal to sub-seasonal forecasting.

White: How do you think that’s helpful?

Lazear: That will be very helpful. If you’re able to say to someone, “There is an increasing likelihood over the next month or two of potentially major flooding,” or, “The pattern looks favorable for major flooding in New York State,” that can be really helpful to help allocate resources and spread the word.

White: In the writing world, there is a great deal of skepticism about artificial intelligence and what it means for the profession. Did you hear any objection to AI from attendees in Baltimore?

Lazear: I think people in the atmospheric science community are excited about where this is going. I was also happy to hear people at the National Weather Service and government recognize that this is the future. We have to change the way we work to embrace this and improve the ways we disseminate weather forecasts to the public. This is what we have to do in order to continue to make ourselves relevant.

White: Does this mean forecasters have to adapt their roles?

Lazear: I think the role of humans as forecasters will become more about communicating with the right people at the right time, unlike 20 years ago when you had to train a forecaster to make accurate forecasts of high and low temperatures. From now on, the role of forecasters can be more as interpreters, coming up with some sort of probabilistic forecast and making sure you communicate that accurately to decision makers. So the role of the forecaster will be less, the highest on Saturday will be 53, and more to whom do we need to communicate this information, when and how? So something like UAlbany’s new State Weather Hazards Communication Center (which aims to improve the state’s preparedness and response to severe weather) is the kind of center where the role of the forecaster becomes increasingly important.

(Tags for translation)State Weather Hazard Communications Center

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *