Hungjui Yu 尤虹叡
Research Scientist at CIRA, Colorado State University, Fort Collins, CO
I study the three-dimensional structure of convective storms and how they evolve across different environments, using satellite and radar observations, numerical models, and machine learning, with applications to high-impact weather and aviation safety.
At CIRA, I lead the development of cloud-free line-of-sight tools for aviation, and collaborate with the Department of Atmospheric Science on algorithms for NASA’s INCUS mission. I am also a co-investigator on NSF- and NOAA-funded projects on AI and storm modes. I received my Ph.D. in Atmospheric Sciences from National Taiwan University and serve as an Associate Editor of Artificial Intelligence for the Earth Systems.
Research
- Weather–Climate Interface – How convective storm modes and heavy rainfall arise from local environments and change in a warming climate.
- Machine Learning for Clouds and Storms – Inferring subgrid-scale convective clouds from coarse-resolution environments to bridge climate models and storm scales.
- Upper-Air Observations & Radiosonde QA/QC – Turning field-campaign soundings (DYNAMO, TASSE, PRECIP) and low-cost radiosondes into research-grade datasets.
- Aviation & Cloud-Free Line of Sight – Quantifying cloud-free line of sight from 3D cloud data to support aviation safety.
Approach
- Observations + models – Combining satellite, radar, and radiosonde observations with convection-permitting simulations.
- Machine learning – Extracting the environmental controls on clouds and storms to improve physical understanding.
- Open-source software – Scalable Python tools and workflows, e.g., Cloud System Classification for convective–stratiform separation and storm-mode classification.