September 29, 2025
SCI is proud to announce that Xiaowei Jia, an associate professor with the Department of Computer Science, has received a prestigious National Science Foundation (NSF) award of $384,203 to advance climate science through machine learning.
With collaborators at the National Center for Atmospheric Research and the University of Maryland, Jia’s research focuses on the hidden signatures in water isotopes, specifically rare forms of hydrogen and oxygen that act as natural recorders of Earth’s climate history. These isotopes are found in archives like ice cores, cave formations, and lake sediments, thus providing insight into past temperature, rainfall, and shifting climate patterns. They also help improve water cycle representations in today’s climate models.
However, incorporating isotopes into global climate models is computationally demanding. This project tackles that challenge by developing a machine-learning-based “emulator” that can rapidly predict water isotope patterns from existing climate simulations. The approach promises to dramatically cut costs and expand access to isotope-enabled modeling.
By combining expertise in climate science, hydrology, and artificial intelligence, the research will:
- Improve understanding of the reason for isotopic variability.
- Enhance model-data comparisons using both modern and paleoclimate observations.
- Support climate data assimilation efforts across the scientific community.
- Provide open-source tools, data, and training resources to benefit researchers, educators, and K-12 outreach programs.
This award underscores NSF’s commitment to advancing both fundamental science and broader societal impact. It also highlights SCI’s role in leading interdisciplinary research at the intersection of AI and Earth system science.