[A-AS01] 高性能スーパーコンピュータを用いた最新の大気科学

[AAS01-P02] Preliminary results of a high-resolution climate simulation using the Non-hydrostatic Icosahedral Atmospheric Model, NICAM, for CMIP6 HighResMIP

*小玉 知央¹、佐藤 正樹^2,1、大野 知紀¹、野田 暁¹、八代 尚³、山田 洋平¹、中野 満寿男¹、清木 達也¹、那須野 智江¹、Chen Ying-Wen²、宮川 知己²、杉 正人⁴、Roh Woosub² (1.独立行政法人海洋研究開発機構、2.東京大学大気海洋研究所、3.理化学研究所計算科学研究機構、4.気象研究所)

キーワード：高解像度気候シミュレーション、全球非静力学モデル、台風

We introduce a series of climate simulations using the Non-hydrostatic Icosahedral Atmospheric Model, NICAM. Though our typical resolution range is 3.5-14 km for seasonal integration, a relatively coarser mesh size is chosen for multi-decadal integration. In Kodama et al. (2015, J. Meteor. Soc. Japan), we have performed the AMIP-type 30-year simulations with a mesh size of 14 km under the present and future boundary conditions. Although we switch off cumulus parameterization scheme in order to keep physics schemes consistently across resolutions between 3.5 and 14 km, the simulated climatology is fairly good, competitive with other climate models. The advantage of a fine-mesh global climate simulation is that atmospheric multi-scale phenomena ranging from large-scale circulation to meso-scale features associated with convection, front, severe rainfall, atmospheric gravity waves are represented in a seamless manner. For example, we discuss statistics of detailed structure of multi-scale convective systems and extremes such as tropical cyclones; Yamada et al. (2017, J. Climate) analyzed the NICAM AMIP-type simulation dataset and showed widening of the intense wind speed area around the tropical cyclones due to global warming. Now, targeting the CMIP6 HighResMIP, we are performing further longer time integrations for 65 years. The simulations are initialized on 1^st January 1950. Because multiple choice of resolution is required, we use mesh sizes of 14, 28, and 56 km. The model used here has been updated and tuned in terms of cloud microphysics (Roh and Satoh, 2014, J. Atmos. Sci.), aerosol, orographic gravity wave and land model to improve performance of the simulated climatology. Of particular interest here is performance in genesis, development, track and structure of tropical cyclone, and we will show some preliminary results including impact of the horizontal resolution on tropical cyclone statistics.