Gan Zhang
University of Illinois Urbana-Champaign, IL·Earth & Environmental Science
Storms such as hurricanes and Nor'easters threaten lives and economies, yet some of the most dangerous ones remain hard to predict. Two linked questions run through this research.
- Fields
Earth & Environmental Science
Computer Science
- Topics
tropical cyclones
hurricanes
Nor'easters
- Grants
- 2 known grants, through Dec 2028
What they work on
- Map the limits of how far ahead high-impact tropical and mid-latitude cyclones can be predicted, using new AI weather forecasting models, and identify the starting conditions that control storm development.
- Adapt representation-learning AI algorithms to automatically flag atmospheric precursors, the early warning signs that influence long-range cyclone forecasts.
- Use AI and statistical models together with physical and physics-informed simulations to test how ocean, sea ice and land conditions shape cyclone evolution and predictability.
- Test how the Hadley and Walker overturning circulations set tropical cyclone frequency under greenhouse warming, using aquaplanet and ridge versions of the Community Earth System Model (CESM).
- Run high-resolution (0.25 degree) slab-ocean simulations that reproduce aquaplanet and ridge sea surface temperature patterns, to capture tropical cyclone behavior, and release the model setups to the research community.
More about the research
The first is how far ahead high-impact storms can be forecast. The second is how the wider climate system, including ocean temperatures, large-scale air circulation and greenhouse warming, shapes how many storms form and how they behave. The work combines artificial intelligence (AI) forecasting tools with simplified climate model experiments.
Why it matters
Better long-range warnings of dangerous hurricanes and winter storms could give coastal and inland communities more time to prepare, potentially saving lives and reducing economic damage. A clearer understanding of how greenhouse warming changes tropical cyclone frequency could help forecasters, planners and insurers assess future storm risk. Openly shared AI tools and climate model setups could speed up improvements in weather and climate prediction.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Collaborative Research: CAIG: Unraveling the Long-Range Predictability and Environmental Dependency of High-impact Cyclones with Artificial Intelligence Tools
NSF · Jan 2026 – Dec 2028
Collaborative Research: Understanding the Links between Tropical Cyclones and Tropical Circulation under Climate Change through Idealized Coupled Climate Modeling
NSF · Nov 2023 – Oct 2027
Closest labs
Haiyan Jiang
Florida International University, FL·Earth & Environmental Science
Characterize how tropical cyclone rainfall distribution and asymmetry change before, during and after landfall, using statistical analysis of long-term satellite observations.
Lantao Sun
Colorado State University-Fort Collins, CO·Earth & Environmental Science
Determine how much Arctic sea ice loss affects summer atmospheric circulation and the jet stream, and through what mechanisms, using DAMIP and PAMIP climate model simulations.
Qinghua Ding
University of California-Santa Barbara, CA·Earth & Environmental Science
Develop regional models of summertime Arctic sea ice that combine global observational data through machine learning, to forecast ice from weeks to a season ahead and find how far ahead prediction is possible.
The school
University of Illinois Urbana-Champaign: 41 labs on PI Pages, $907M a year on research, 870 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.