Lightning is one of the most destructive forces of nature, as in 2020 when it sparked the massive California Lightning Complex fires, but it remains hard to predict. A new study led by the 91̽�� shows that machine learning — computer algorithms that improve themselves without direct programming by humans — can be used to improve lightning forecasts.

Better lightning forecasts could help to prepare for potential wildfires, improve safety warnings for lightning and create more accurate long-range climate models.

“The best subjects for machine learning are things that we don’t fully understand. And what is something in the atmospheric sciences field that remains poorly understood? Lightning,” said , a 91̽��associate professor of atmospheric sciences. “To our knowledge, our work is the first to demonstrate that machine learning algorithms can work for lightning.”

The new technique combines weather forecasts with a machine learning equation based on analyses of past lightning events. The hybrid method, Dec. 13 at the American Geophysical Union’s fall meeting, can forecast lightning over the southeastern U.S. two days earlier than the leading existing technique.

“This demonstrates that forecasts of severe weather systems, such as thunderstorms, can be improved by using methods based on machine learning,” said , who did the work for his 91̽��doctorate in atmospheric sciences. “It encourages the exploration of machine learning methods for other types of severe weather forecasts, such as tornadoes or hailstorms.”

Researchers trained the system with lightning data from 2010 to 2016, letting the computer discover relationships between weather variables and lightning bolts. Then they tested the technique on weather from 2017 to 2019, comparing the AI-supported technique and an existing physics-based method, using actual lightning observations to evaluate both.

The new method was able to forecast lightning with the same skill about two days earlier than the leading technique in places, like the southeastern U.S., that get a lot of lightning. Because the method was trained on the entire U.S., its performance wasn’t as accurate for places where lightning is less common.

The approach used for comparison was a recently developed technique to forecast lightning based on the amount of precipitation and the ascent speed of storm clouds. That method has projected and a continued .

“The existing method just multiplies two variables. That comes from a human’s idea, it’s simple. But it’s not necessarily the best way to use these two variables to predict lightning,” Kim said.

The machine learning was trained on lightning observations from the , a collaborative based at the 91̽��that has tracked global lightning since 2008.

“Machine learning requires a lot of data — that’s one of the necessary conditions for a machine learning algorithm to do some valuable things,” Kim said. “Five years ago, this would not have been possible because we did not have enough data, even from WWLLN.”

Commercial networks of instruments to monitor lightning now exist in the U.S., and newer geostationary satellites can monitor one area continuously from space, supplying the precise lightning data to make more machine learning possible.

“The key factors are the amount and the quality of the data, which are exactly what WWLLN can provide us,” Cheng said. “As machine learning techniques advance, having an accurate and reliable lightning observation dataset will be increasingly important.”

The researchers hope to improve their method using more data sources, more weather variables and more sophisticated techniques. They would like to improve predictions of particular situations like dry lightning, or lightning without rainfall, since these are especially dangerous for wildfires.

Researchers believe their method could also be applied to longer-range projections. Longer-range trends are important partly because lightning affects air chemistry, so predicting lightning leads to better climate models.

“In atmospheric sciences, as in other sciences, some people are still skeptical about the use of machine learning algorithms — because as scientists, we don’t trust something we don’t understand,” Kim said. “I was one of the skeptics, but after seeing the results in this and other studies, I am convinced.”

Other collaborators are and at the UW, and Yoo-Geun Ham and Jeong-Hwan Kim at Chonnam National University in South Korea.

For more information, contact Kim at daehyun@uw.edu or Cheng at wycheng@uw.edu. Cheng will this research online at 12:45 p.m. Central Time (10:45 a.m. Pacific, 1:45 p.m. Eastern) on Monday, Dec. 13.

91̽�� scientists have published a mathematical model that could help explain and forecast the Madden-Julian Oscillation, a massive cluster of thunderstorms that plays a role in global weather.

“Over the Indian Ocean and the Western Pacific – one of the warmest, most moist areas of the planet – there is a colossal cluster of clouds every 40 to 50 days or so,” said corresponding author , a 91̽��doctoral student in atmospheric sciences. “When it’s active, it’s a very strong signal.”

The , as it is known, is a gargantuan cluster of rain clouds that pummels the Earth below it, bringing a moving collection of rainstorms that lasts a week or more, followed by a period of mostly clear skies, and repeats about every month and a half. Since people’s memory of weather from more than a month ago tends to be fuzzy, the pattern wasn’t discovered until weather balloons showed its existence in the 1970s.

Today, satellite observations can clearly show storms moving across the tropical Pacific every 45 days or so, before a quiet phase of about a month or more allows the system to rebuild. Several such cycles repeat, and then the system enters a quiet phase when it can be inactive for many months.

“We know what the MJO is doing now, but we have trouble knowing what it’s going to do next,” Adames said.

His new , to be published in the Journal of the Atmospheric Sciences, presents a series of equations that describe how the cluster of storms moves, and where the next will pop up.

“We believe MJO will be the next El Niño,” said co-author , a 91̽��assistant professor of atmospheric sciences.

Better understanding of the MJO would help predict tropical rainstorms and flooding over India, northern Australia, and Pacific islands such as the Maldives and Indonesia. It could also improve medium-range global forecasts, since the MJO can nudge weather patterns that affect the mainland U.S. For example, the MJO can amplify the effects of El Niño, as it did in early January 2016, when both systems were active in the tropical Pacific at the same time.

While MJO storm clouds move eastward, the equations in the new study predict that, in the weeks that follow, clear sky conditions will develop to the west of where the storm system started, which in turn will cause the development of another region of storminess even further west.

“We came up with a theoretical model that can explain almost all of the fundamental features of the MJO,” Kim said.

“This is an idealized model, so it won’t help with forecasting yet,” Adames said. “But it shows that we’re beginning to understand the physics behind this system.”

The model explains three major elements of the MJO: Its huge size, its timescale of about 45 days, and why the storm clouds always move toward the east.

To do so, the model combines several recent theories about the MJO. One is the idea that unlike places such as the continental U.S. during winter, where the interaction between cold and warm air masses drive precipitation, in the tropics humidity reigns. The model asserts that the MJO creates high-humidity areas which eventually develop into large thunderstorms.

“I’m from Puerto Rico, and I can say with certainty that there are fewer contrasts in temperature in the tropics,” Adames said. “Another way to drive the weather is by changing the amount of humidity in the atmosphere. This is how the MJO changes weather conditions. ”

The model also incorporates the idea that as the moisture wave we see travels east, it disturbs the atmosphere to its west, setting up the stage for the next cluster of storm clouds.

The researchers used satellite observations from NASA’s to check their math. Indeed, the data show that each storm begins slightly to the west of the preceding one, and matches the patterns they predicted.

Still mysterious is what first sets up an MJO. It tends to be more active in winter, but sometimes disappears for months at a time. Also unknown is why some MJOs reach the Pacific, where they affect global weather, while others just peter out in the Indian Ocean. These will be areas of future study, Kim said, as well as looking at how to improve representation of the MJO in global climate models.

The research was funded by NASA.

###

��

For more information, contact Adames at angelf88@atmos.washington.edu and Kim at daehyun@uw.edu or 206-221-8935.

��

NASA grant number: NX13AM18G