Researchers from Japan’s RIKEN Advanced Institute for Computational Science visited the University of Maryland’s Earth System Science Interdisciplinary Center (ESSIC) in October to compare notes on data assimilation methods and capabilities.
The process of incorporating data into predictive models requires a multitalented team and plenty of processing power, said UMD Associate Professor Kayo Ide, who worked with ESSIC Director Antonio Busalacchi and RIKEN’s Takemasa Miyoshi to organize the conference.
UMD and RIKEN boast different strengths when it comes to data assimilation. In College Park, university researchers have better government connections, facilitating collaboration with NOAA, NASA and other agencies, Ide said. But RIKEN has a supercomputer than can run 100 times as many simulations as UMD’s – “It’s no comparison.”
Representatives of the two institutions met at ESSIC on Oct. 7-9 in the hopes of laying groundwork for future collaboration.
“UMD is one of the world's top centers on data assimilation research,” Miyoshi wrote in an email. He earned his Ph.D. in meteorology on ensemble data assimilation from UMD and worked at the university as a research assistant professor and assistant professor before becoming the leader of RIKEN’s Data Assimilation Research Team (DART).
“In RIKEN, we started my group on data assimilation research only three years ago, and are still spinning up,” Miyoshi wrote. “We have a lot to learn from UMD, such as math-weather collaboration.”
Ide, a member of UMD’s Weather-Chaos Group, is working to develop a space weather forecast – a model for conditions in the upper atmosphere. She is also involved with the Joint Center for Satellite Data Assimilation – which includes NOAA, NASA, the U.S. Air Force and Navy, and other groups – in coordinating satellites to prevent gaps in orbital data.
The meeting with RIKEN researchers was a “bottom-up approach” to coordinating research, Ide said, fostering personal connections that can spur future work.
RIKEN postdoctoral researchers Takumi Honda and Shunji Kotsuki appreciated the opportunity to meet “famous professors” as well as young researchers, they both wrote in emails.
“We were able to feel the atmosphere of the world-leading University and meet the people there,” wrote Honda, who also appreciated the group’s visit to the NOAA Center for Weather and Climate Prediction, located near ESSIC.
Ide noted that promoting communication among UMD researchers is important, too, as data assimilation for environmental study is an interdisciplinary effort involving engineers and mathematicians as well as atmospheric and oceanic scientists.
“Just getting to know who’s on the campus is difficult,” Ide said.
At the same time, she added, the applications of data assimilation are growing. She noted, for example, that data about cancer patients could be processed to forecast new patients’ outcomes and improve treatment.
“Anything you can get data [on], and any system that can make a prediction, data assimilation is applicable,” Ide said.