Big Data Assimilation

15-2 January 24 14:30-14:45

Current status of SCALE and its future toward the accurate numerical weather prediction

Hirofumi Tomita (RIKEN), Seiya Nishizawa (RIKEN), Tsuyoshi Yamaura (RIKEN), Ryuji Yoshida (RIKEN), Hisashi Yashiro (RIKEN), Kenta Sueki (RIKEN)


During the big data assimilation project, our team has been developing a high-resolution atmospheric regional model suitable to capture each deep convection with high computational efficiency. Due to limitation of allocation time for simulation and, we improved the model with several remedies, nesting model framework and employment of single-precision. For the improvement of physical performance, we improved the scheme for steep mountain, surface flux scheme, advection, and so on. The turbulence scheme in the current version is based on the Raynolds Averaged Navier-Stokes (RANS) model because the horizontal grid spacing is several 100m. To do more accurate simulation by fully utilizing data of the phased array weather radar, the turbulence scheme should be Large Eddy Simulation (LES) model that is more principle way than RANS with higher resolution. However, there are still issues for LES schemes used in the meteorology: numerical convergence problem for statistics deep convection and theory associated with moist process. After review of model improvement in this period, we talk about such future issues toward more accurate real-time data assimilation system from the viewpoint of model construction.

Contact information

Dr. Hirofumi Tomita RIKEN