High Performance Computing for Mesoscale Weather Prediction with the K-computer

A research on super high-resolution mesoscale numerical weather prediction with the High Performance Computing of the K-computer is introduced. This research is one of the five fields of the MEXT-funded national research project in Japan, the HPCI Strategic Programs for Innovative Research (SPIRE). Following three subjects are underway to show the feasibility of precise prediction of local high impact weather phenomena using HPC: 1) Development of cloud resolving 4-dimensional data assimilation systems, 2) Development and validation of a cloud resolving ensemble analysis and forecast system, and 3) Basic research with very high resolution atmospheric models. In the presentation, development of advanced data assimilation methods (e.g., LETKF and hybrid 4D-VAR) is introduced, and examples of high-resolution and super high-resolution data assimilation/ensemble experiments for mesoscale high impact weathers (torrential rains, tropical cyclones, and tornados) are shown with their visualization.

Key words: High Performance Computing, K-computer, Mesoscale Numerical Weather Prediction