Natural Hazards and Earth System Modeling

We develop novel GPU-enabled high-performance computing frameworks, reduced-order models, and data assimilation techniques to understand and predict complex Earth system processes and natural hazards——including landslides, ice sheet dynamics, flooding, wildfire, and more. Our research places particular emphasis on thermal-hydro-mechanical (THM) coupling in geosystems, especially under extreme environmental conditions.

Dam break simulation by neural meshfree method

Natural Hazards

Landslides

Landslide Modeling

Collapse of granular column

 

collapse of granular column

Compound Flooding Modeling

PINN for SVE

Relavant Papers:

  • Feng, D., Tan, Z., He, Q. (2023) Physics-informed neural networks of the Saint-Venant equations for flood modeling. Water Resources Research.
     

Glaciers

Ice Sheet Modeling

Greenland
Hybrid model for ice sheet modeling
  • He, Q., Perego, M., Howard, A. A., Karniadakis, G. E., & Stinis, P. (2023). A hybrid deep neural operator/finite element method for ice-sheet modeling. Journal of Computational Physics, 492, 112428.