AI-NWP International Workshop: Advancing Taiwan–U.S. Collaboration in Weather Forecasting Technologies

The AI-NWP International Workshop was held by NCHC in Hsinchu on February 17, 2025, focusing on the application of artificial intelligence (AI) and GPU acceleration technologies in Numerical Weather Prediction (NWP). The event aimed to foster technical collaboration between Taiwan and the U.S. National Center for Atmospheric Research (NCAR) to jointly improve the performance and accuracy of weather simulations.


The workshop brought together research teams from Taiwan’s Central Weather Administration (CWA), Academia Sinica, National Taiwan University, National Yang Ming Chiao Tung University, and the National Center for High-performance Computing (NCHC), who engaged in in-depth discussions with NCAR experts. Through this international collaboration, Taiwan’s research institutions can benefit from NCAR’s extensive experience in high-performance computing (HPC) and AI, thereby enhancing the precision and efficiency of local weather models and laying a solid technical foundation for forecasting and responding to extreme climate events.
Experts from both Taiwan and the U.S. delivered a series of keynote presentations exploring how AI and GPU acceleration technologies can improve weather forecasting capabilities. Highlights included:
•    Dr. Thomas Hauser, Deputy Director of NCAR and Director of the Computational and Information Systems Laboratory (CISL), presented “Overview of Our Data and Infrastructure Strategy for AI and GPU Acceleration,” analyzing the potential of AI in weather simulation.
•    Sheri Voelz (NCAR) introduced “Streamlining Software Development for MPAS-A: A New Workflow Strategy,” outlining approaches to simplify atmospheric model development.
•    John Dennis (NCAR) shared “Experiences with Performance Portability of Several Atmospheric Physics Modules,” offering insights on optimizing physics models across platforms.
On the Taiwan side:
•    Dr. Cheng-Chin Liu of the Central Weather Administration gave a talk on “The Current Status and Future Planning of AI/ML and GPU Acceleration in Meteorological Applications,” demonstrating how AI is enhancing Taiwan’s weather forecasting capabilities.
•    Deputy Director Chih-Huang Hsiao of NCHC introduced “HPC Infrastructure at NCHC,” showcasing national-level computing resources.
•    Dr. Huang-Hsiung Hsu from Academia Sinica’s Research Center for Environmental Changes (RCEC) shared work on “Applying GPU Power to Climate Change Models,” detailing RCEC’s early-stage AI and acceleration efforts.
•    Prof. Hung-Chi Kuo of National Taiwan University presented “AI Insights for Meteorological Models and Taiwan's Data-Driven Weather Prediction Development,” highlighting innovation in AI-based forecasting.
•    Prof. Hsin-Mu Tseng of National Yang Ming Chiao Tung University delivered “AI-Based Downscaling for Meteorology,” emphasizing AI’s role in regional weather prediction.
This workshop underscored the transformative potential of AI and HPC in the meteorological field and set the stage for deeper Taiwan–U.S. cooperation. It is anticipated that continued collaboration will accelerate the integration of AI technologies into weather forecasting, offering robust technological support for addressing global climate change challenges.



NCHC Director General Chau-Lyan Chang delivers opening remarks


CWA Section Chief Po-Liang Chang opens the workshop
 
 Dr. Thomas Hauser presents “Overview of Our Data and Infrastructure Strategy for AI and GPU Acceleration”


Sheri Voelz (NCAR) introduces “Streamlining Software Development for MPAS-A: A New Workflow Strategy”
 

John Dennis (NCAR) shares “Experiences with Performance Portability of Several Atmospheric Physics Modules”