Speeding through the stratosphere, some twenty-two miles above the Earth, the United States’ next generation weather satellite watched as storm clouds churned over the Midwest.
On the ground, more than 200 tornadoes from New Jersey to Texas devastated communities during a two-week period in May 2019, marking one of the largest tornadic outbreaks in recent history.
High above the storms, the Geostationary Operational Environmental Satellite-16 (GOES-16) captured dramatic, high-resolution images every thirty seconds and beamed the information back to Earth.
Data like these could someday improve our understanding of severe weather and increase warning times, but only if scientists can help forecast models ingest the massive amounts of information.
Back on Earth, researchers with Penn State’s Department of Meteorology and Atmospheric Science and the Center for Advanced Data Assimilation and Predictability Techniques (ADAPT) are doing just that.
Fuqing Zhang, a distinguished professor of meteorology and a world leader in the field of data assimilation, led the ADAPT team before his death in July 2019. However, Penn State researchers are continuing the important work of creating cutting-edge algorithms that can feed data from the newest satellites into forecast models that predict daily weather, warn of storms and health hazards, and inform farmers when to plant and harvest their crops.
“Data assimilation uses statistics to combine high technology observations with the existing high-resolution numerical weather prediction models,” Zhang said in an interview. “We integrate the observations with the models to help them do the best job.”
Improving warning times
About forty years ago, people in the path of potentially deadly storms like those that ripped through the United States in May could expect little warning—three minutes on average. That number is about fourteen minutes today, thanks to improvements in numerical weather prediction.
Forecasting models better simulate the physics behind prevailing westerlies, storms and fronts that make up Earth’s climate, and new satellites provide better images than ever. Advanced data assimilation techniques, like those Zhang and his laboratory have developed, help combine these models and data.
“If you want to get weather right, you need to know what is going on now, and what has been occurring recently,” said Richard Alley, Evan Pugh University Professor of Geosciences, who has collaborated with Zhang on previous research.
Data assimilation helps feed forecast models the most accurate picture of current conditions, important because even small changes to the atmosphere can lead to divergent forecasts after time, a phenomenon sometimes called the “butterfly effect.”
“This requires lots of data and it requires techniques to tell the computer models what the data say,” Alley said. “Fuqing’s work has been central in improving the ability to put data and models together to give better forecasts. This is not easy or obvious to do; there is a real art in the science of how to put data and models together to get more value out of both.”
In October 2018, Zhang and his team became the first to use data obtained from GOES-16 in a numerical weather prediction model used to provide guidance for tornadic thunderstorm forecasting. The results suggest more advanced warnings are possible.
For the tornado research, Zhang worked with David Stensrud, professor and head of the Department of Meteorology and Atmospheric Science, and Yunji Zhang, assistant research professor in meteorology. In another paper published in November, Yunji Zhang and Stensrud found combining data from GOES-16 with traditional weather radar produced even better forecasts that maintain the early warning times and produce more accurate predictions.
“Say you have severe weather heading toward a football game or a large event,” Yunji Zhang said. “If you can have a longer forecast lead time of twenty to forty minutes, you have more time to evacuate. I believe that more human lives can be saved by increasing forecast times.”
Parked in geostationary orbit high above the Pacific Ocean, the GOES-17, the newest U.S. weather satellite, started taking breathtaking new images in Earth after going fully operational in February 2019.
The satellite, which along with the GOES-16 monitors the entire country and much of the western hemisphere, uses multiple bands of visible and infrared light to reveal factors such as fog, winds, vegetation, snow and ice, fires, water vapor, and lightning.
Traditional forecast models, however, have only been able to use clear-sky observations, which ignore images with cloudy or rainy skies due to challenges in assimilating the information.
Fuqing Zhang and members of his laboratory developed a method to incorporate all-sky radiance data into forecast models.
“All-sky means using data from both clear sky and rainy, cloudy conditions. And, by the way, that’s usually where the severe weather actually occurs,” Zhang said. “Our team is the first in the world to directly assimilate the all-sky radiance from a geostationary satellite with cloud-resolving weather prediction models.”
The all-sky radiance technique also shows promise for forecasting longer-scale weather events, snowstorms, and hurricanes.
Using the method, Zhang and his team correctly showed Hurricane Harvey would become a Category 4 storm while existing models forecast it as a Category 1. The results are published in the July 2019 issue of the Bulletin of the American Meteorological Society.
“We are especially proud of how we can assimilate the inner circulation of the eye wall, much better than anything before,” Zhang said. “And we think this has great potential in improving hurricane forecasts.”
The College of Earth and Mineral Sciences mourns the death of professor Fuqing Zhang who died on July 19, 2019, not long after being diagnosed with cancer. He was 49.
Zhang will be remembered for his energy, enthusiasm, good sense of humor and love of life. He also will be remembered for his many contributions to his field, particularly in data assimilation and prediction science; for the numerous early-career scientists he mentored; and for the myriad scientists from other disciplines whom he introduced to data assimilation.
“Fuqing’s pioneering data assimilation and predictability research has vastly improved our ability to accurately predict hurricanes and other severe weather phenomena,” said David Stensrud, head of the Department of Meteorology and Atmospheric Science.
Zhang earned his bachelor of science degree in 1991 and master of science degree in 1994, both in meteorology, from Nanjing University, China. He earned his doctorate in atmospheric science in 2000 from North Carolina State University. https://bit.ly/2RMsqi7