The expansive whiteness of the ice and snow at the Antarctic seems the same as more than a century ago when explorers first started traversing the continent. But scientists monitoring the continent have noticed rapid changes in the mass and reach of ice sheets, whose collapse can impact sea levels all over the world. Climate change has raised questions about the physics of these ice sheets and how they are changing the polar region.
The research of Ching-Yao Lai seeks to answer these pressing questions. Lai is an assistant professor of geosciences and atmospheric and oceanic sciences and is a participating faculty member of the Center for Statistics and Machine Learning (CSML). Her research – lying at the intersection of climate science, geophysics, and fluid dynamics – uses observational data with statistical and machine learning models to understand ice sheet dynamics.
“I am a modeler. I develop mathematical descriptions to understand how much ice will be lost in the next century as the climate warms,” said Lai, summing up her research focus. “The current source for sea level increases is the melting of the Antarctic and Greenland ice sheets. There are many physical processes that govern how fast these ice sheets are melting. In my research, I'm trying to develop models to understand and make better predictions of the ice loss.”
Scientists have been closely examining and calculating ice sheet mass and dynamics for decades with the help of satellite imagery, radar, measurements from aircraft, and observations and measurements from the field. Lai takes a slightly different approach.
“The novel part of my work that is very different from the rest of the people in my community is that we are actively trying to push machine learning, especially deep learning, in order to improve our traditional way of understanding ice sheet dynamics.”
For example, Lai has been looking at iceberg calving, when ice sheets lose mass as huge chunks of ice break off ice sheets or glaciers. How do you predict these ice-calving events? This is a complex question to answer because ice sheets are dynamic bodies that are pushed and pulled by different factors such as temperature and geography.
Satellite imagery shows the surface of an ice sheet and shows fractured patterns. These cracks can develop into icebergs, said Lai. She and her research group look at thousands of satellite images of these cracks and train a convolutional neural network to identify the location of these cracks. The machine learning model takes this input and predicts the location of future cracks. Traditional methods of mapping Antarctic cracks are slower and cannot be scaled up to an entire continent. Machine learning can do predictions on a larger geographical scale and do so more quickly than older methods, Lai said. Through the machine learning approach, she produced the first Antarctic-wide map of fractures.
Lai has been an assistant professor at Princeton since 2021. Besides the geosciences department and the program in atmospheric and oceanic sciences and CSML, Lai is also affiliated with Princeton’s High Meadows Environmental Institute.
Before coming to Princeton, Lai was a Lamont Postdoctoral Fellow at the Lamont-Doherty Earth Observatory at Columbia University from 2018 to 2020. She earned her doctoral degree in mechanical and aerospace engineering in 2018 at Princeton, advised by notable fluid dynamics expert, Howard Stone, the Donald R. Dixon '69 and Elizabeth W. Dixon Professor of Mechanical and Aerospace Engineering and chair of the Department of Mechanical and Aerospace Engineering.
During her doctoral years, Lai earned the Maeder Graduate Fellowship in Energy and the Environment for her prolific research accomplishments, such as three first-author papers in the journals Physical Review Letters; Philosophical Transactions of the Royal Society A: Mathematical, Physical & Engineering Sciences; and Proceedings of the Royal Society A: Mathematical, Physical & Engineering Sciences. Her fellowship was spent studying hydraulic fracturing, a technique to extract fossil fuels using high-pressure liquid from shale formations. Her fellowship specifically focused on the use of foams to decrease water use in hydraulic fracturing, an environmental concern.
Lai received her bachelor’s degree in physics from National Taiwan University in Taipei, Taiwan in 2013.
Thus far in her career as an early-career researcher, Lai has published more than 15 research papers and given more than 40 talks at various conferences and institutions such as the Massachusetts Institute of Technology and the University of Oxford.
“What I find exciting about my research and the work my research group is doing is the machine learning component,” said Lai. “We are adding deeper understanding to a natural phenomenon that may impact policy and ultimately how we confront climate change in the future.”
