Articles | Volume 30, issue 21
https://doi.org/10.5194/hess-30-6843-2026
https://doi.org/10.5194/hess-30-6843-2026
Research article
 | 
09 Nov 2026
Research article |  | 09 Nov 2026

A physically integrated GNN surrogate for microbially-mediated kinetic reactive transport: PHYGNET

Jinbo Wang, Kunfeng Zhang, Walter Illman, Shuai Chen, and Mingzhu Liu

Model code and software

Code and data for PHYGNET (Version v1.0) J. Wang https://doi.org/10.5281/zenodo.17797476

Download
Short summary
PHYGNET, a physical AI model, designed to accelerate microbial reactive transport simulations. This new approach is over 3500 times faster than traditional methods and handles complex, large-scale simulations without crashing. By allowing "small-sample training, ultra-large-scale inference", it cuts weeks of environmental modeling work down to hours, making it practical for real-world cleanup and risk assessment projects.
Share
Altmetrics
Final-revised paper
Preprint