Annual Report of the Earth Simulator Center April March 2013 The Earth Simulator Center, Japan Agency for Marine-Earth Science and Technology

Transcription

1 ISSN

2 Annual Report of the Earth Simulator Center April 2012 March 2013

3

4 O CONTENTS utline of th e Earth Simulator P roj ect 1. Mission and Basic Principles of the Earth Simulator 1 2. Earth Simulator Research Project 1 3. Collaboration Projects 4. System onfiguration of the Earth Simulator 4 Earth Simulator Research P roj ects C h apter 1 Earth Science nderstanding oles of ceanic Fine Structures in limate and Its ariability I 11 Wataru hfuchi Earth Simulator enter, apan gency for Marine-Earth Science and Technology 海 洋 の 前 とそれらが 生 出 す 大 海 洋 現 の 解 明 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター 大 daptation riented Simulations for limate ariability 17 eiko Takahashi Earth Simulator enter, apan gency for Marine-Earth Science and Technology 変 に 適 可 能 な 環 境 探 のためのマルチス ールシミュレーション 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター 高 Development of a igh- esolution oupled limate Model for Global Warming Studies 25 kira oda esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 地 球 化 研 究 のための 高 度 モデルの 開 発 研 究 海 洋 研 究 開 発 機 構 地 球 環 境 変 野 Simulations of tmospheric General irculations of Earth-like lanets by FES 31 Yoshi-Yuki ayashi Department of Earth and lanetary Sciences, obe niversity AFES を 用 いた 地 球 の 大 大 環 シミュレーション 大 大 研 究 Study on the Diagnostics and rojection of Ecosystem hange ssociated with Global hange 39 Michio. ishi esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 地 球 環 境 変 化 に う 生 変 の 断 と に 関 する 研 究 海 洋 研 究 開 発 機 構 地 球 環 境 変 郎 FD Simulation of Summertime ir Temperature Distribution of the Twenty ilo-meters S uare rea of enter of ong ong 43 Yasunobu shie ational Institute for and and Infrastructure Management CFD による 20km の 分 シミュレーション 技 術 研 究 信 Study of loud and recipitation rocesses sing a Global loud esolving Model 47 Masaki Satoh esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology tmosphere and cean esearch Institute, The niversity of Tokyo 球 解 モデルを 用 いた プロセス 研 究 海 洋 研 究 開 発 機 構 地 球 環 境 変 大 大 海 洋 研 究 i

5 Seasonal rediction Experiment and rocess Studies sing oupled General irculation Model 55 Yukio Masumoto esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 大 海 洋 結 モデルを 用 いたプロセス 研 究 と 実 験 海 洋 研 究 開 発 機 構 地 球 環 境 変 Simulation and erification of Tropical Deep onvective louds using Eddy- ermitting egional tmospheric Models II 61 ozo akamura esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 解 可 能 な 大 モデルを 用 いた い 対 流 のシミュレーションとその (その ) 海 洋 研 究 開 発 機 構 地 球 環 境 変 中 Data Synthesis for iogeochemical ariables by sing a Dimensional ariational pproach 67 Shuhei Masuda esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 次 元 変 分 法 を 用 いた 生 化 変 観 データの 海 洋 研 究 開 発 機 構 地 球 環 境 変 Global Elastic esponse Simulation 71 Seiji Tsuboi Data esearch enter for Marine-Earth Sciences, apan gency for Marine-Earth Science and Technology 地 球 性 シミュレーション 海 洋 研 究 開 発 機 構 地 球 報 研 究 センター umerical Simulations for Electromagnetic rocesses in the Earth 77 Yozo amano Institute for esearch on Earth Evolution, apan gency for Marine-Earth Science and Technology 地 球 内 の に 関 する 数 シミュレーション 海 洋 研 究 開 発 機 構 地 球 内 イ ミクス 野 洋 Development of a Distinct ater hase ssociated with Shallow, uter-rise Earth uakes along the apan Trench - FDM Simulation using the Earth Simulator 83 Takashi Furumura enter for Integrated Disaster Information esearch, Interfaculty Initiative in Information Studies, The niversity of Tokyo / Earth uake esearch Institute, The niversity of Tokyo 海 いの 地 の の 特 異 な についての 地 球 シミュレータによる シミュレーション 大 大 報 環 報 研 究 センター 大 地 研 究 Development of Advanced Simulation Methods for Solid Earth Simulations 91 Mikito Furuichi Institute for esearch on Earth Evolution, apan gency for Marine-Earth Science and Technology 的 体 地 球 シミュレーションコードの 開 発 海 洋 研 究 開 発 機 構 地 球 内 イ ミクス 3D umerical Simulations of Eruption louds ase Study of the 2011 Shinmoe-dake Eruptions 97 Takehiro oyaguchi Earth uake esearch Institute, The niversity of Tokyo の 次 元 数 シミュレーション 2011 年 の 再 現 大 地 研 究 Space and Earth System Modeling 103 anya usano aboratory for Earth Systems Science, apan gency for Marine-Earth Science and Technology 地 球 表 地 球 内 の 関 モデリング 海 洋 研 究 開 発 機 構 システム 地 球 ラボ 野 ii

6 umerical Experiments with Multi-Models for aleo-environmental roblems 109 yako be- uchi tmosphere and cean esearch Institute, The niversity of Tokyo 環 境 研 究 のための 数 実 験 大 大 海 洋 研 究 Development of a igh- esolution limate Model for Model- bservation Integrating Studies from the Earth s Surface to the ower Thermosphere 117 Shingo Watanabe esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 高 解 度 モデルの 開 発 地 表 から 下 大 のモデル 観 研 究 に て 海 洋 研 究 開 発 機 構 地 球 環 境 変 redictability ariation in umerical Weather rediction a Multi-Model Multi- nalysis pproach 123 Takeshi Enomoto Disaster revention esearch Institute, yoto niversity 数 報 にお る 可 能 性 変 カニ ムの 解 明 大 研 究 esearch on the limate hange Mechanisms in the rctic egions ased on Improvements of the old- egion rocesses and alidations of the rctic limate eproducibility sing Global limate Models 129 Yoshiki omuro esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology 球 モデルを 用 いた プロセス 高 度 化 と 再 現 性 による 変 カニ ムの 研 究 海 洋 研 究 開 発 機 構 地 球 環 境 変 C h apter 2 Epoch -Making Simulation arge Scale Simulations for arbon anotubes 137 Satoshi akamura esearch rganization for Information Science Technology カーボン チューブの 特 性 に 関 する 大 シミュレーション 高 度 報 技 術 研 究 機 構 中 arge-scale Simulation for a Terahertz esonance Superconductor Device 143 Mikio Iizuka esearch rganization for Information Science and Technology テラ ル 発 に 関 する 大 シミュレーション 高 度 報 技 術 研 究 機 構 Direct umerical Simulations of Fundamental Turbulent Flows with the World s argest umber of Grid-points and pplication to Modeling of Engineering Turbulent Flows 149 Takashi Ishihara Graduate School of Engineering, agoya niversity 流 の 最 大 数 計 算 とモデリングによる 用 計 算 大 大 研 究 原 arge-scale Genomics and roteomics nalyses onducted by the Earth Simulator 155 Toshimichi Ikemura agahama Institute of io-science and Technology ム タン ク の 化 マップを 用 いた 大 スト ム 解 析 バイオ 大 バイオ イ ンス iii

7 Development of a Fluid Simulation pproach by Massively arallel it-wise perations with a ew iscosity ontrol Method 161 iroshi Matsuoka esearch rganization of Electrical ommunication, Tohoku niversity 性 法 による ビット 算 流 体 シミュレーション 法 の 開 発 大 信 研 究 松 岡 Evaluation of the Intensity of 1,02 -bit S ryptography ode 167 idehiko asegawa Faculty of ibrary, Information and Media Science, niversity of Tsukuba 1024 ビット RSA の 強 度 定 大 報 ディア 川 arge-scale Electronic Structure alculations of iomolecular and elated Systems by the Fragment Molecular rbital Method 173 Shigenori Tanaka Graduate School of System Informatics, obe niversity フラグ ント 分 法 による 生 体 分 に 対 する 大 計 算 大 大 システム 報 研 究 中 成 Evaluations of Steady and nsteady lood essel Wall Stresses of an rtery with a erebral neurysm 179 Tadashi Tanuma pplied Fluid Dynamics Energy Machinery Systems, oint rogram enter, Teikyo niversity を の 定 び 非 定 力 の 評 価 大 ジョイントプログラムセンター Development of the ext-generation omputational Fracture Mechanics Simulator for onstructing Safe and Sustainable Society 187 yuji Shioya Faculty of Information Sciences and rts, Toyo niversity な 可 能 社 会 のための 次 計 算 力 シミュレータの 開 発 洋 大 報 recise alculations for Few- ody tomic Systems using the Gaussian Expansion Method and Applications to Cold Atoms 191 Emiko iyama I E ishina enter for ccelerator- ased Science ウス 関 数 展 開 法 による 数 の 計 算 と 原 への 用 化 研 究 加 器 研 究 センター Strongly onlinear esponse nalyses on Steel Frame Structure with arge-scaled and Super-detailed umerical Modeling 195 Yoichi Mukai Department of rchitecture, Graduate School of Engineering, obe niversity 構 構 の 大 モデリングによる 強 非 解 析 大 大 研 究 洋 一 Development of M D eat Transfer Database at Design onditions of dvanced lanket System in Fusion eactor 201 Yoshinobu Yamamoto Department of Mechanical Systems Engineering, niversity of Yamanashi 進 ブラン ットにお る MHD データ ースの 構 大 大 研 究 機 システム iv

8 C h apter 3 Visualization Studies of arge-scale Data isualization and isual Data Mining 207 Fumiaki raki Earth Simulator enter, apan gency for Marine-Earth Science and Technology 大 データ 可 視 化 とビジュアルデータマイニングの 研 究 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター 荒 木 文 明 v

9

10 Outline of the Earth Simulator Project 1. Mission and Basic Principles of the Earth Simulator The Earth Simulator was developed for the following aims. The first aim is to ensure a bright future for human beings by accurately predicting variable global environment. The second is to contribute to the development of science and technology in the 21st century. Based on these aims, the principles listed below are established for the projects of the Earth Simulator. 1) Each project should be open to researches in each research field and to the public, rather than it is confined within the limited research society. 2) In principle, the research achievements obtained by using the Earth Simulator should be promptly published and returned to the public. 3) Each project should be carried out for peaceful purposes only. 2. Earth Simulator Research Project There are two fields of Earth Simulator Research Projects, as follows: Earth Science Epoch-making Simulation The allocation of Earth Simulator resources for each research field in FY2012 was decided to be as shown in following graph. Public project recruitment for Earth Simulator Research Projects in FY2012 was held in February 2012, and 32 research projects were selected by the Selection Committee. The Allocation of Resources of the Earth Simulator in FY2012 1

11 Authorized Project List in FY2012 Earth Science (20 projects) Project Name Name of Project Representative Professional Affiliation of Project Representative 1 Understanding Roles of Oceanic Fine Structures in Climate and its Variability Wataru Ofuchi ESC, JAMSTEC 2 Simulations of Adaptation-Oriented Strategy for Climate Variability Keiko Takahashi ESC, JAMSTEC 3 Development of a High-quality Climate Model for Global Warming Projection Study Akira Noda RIGC, JAMSTEC 4 Simulations of Atmospheric General Circulations of Earth-like Planets by AFES Yoshiyuki Hayashi Graduate School of Science, Kobe University 5 Study on the Diagnostics and Projection of Marine Ecosystem Change Associated with Global Change Michio Kishi RIGC, JAMSTEC 6 Development of a Numerical Model of Urban Heat Island Yasunobu Ashie National Institute for Land and Infrastructure Management 7 Study of Cloud and Precipitation Processes using a Global Cloud-system Resolving Model Masaki Sato RIGC, JAMSTEC/Atmosphere and Ocean Research Institute, The University of Tokyo Study on the Predictability of Climate Variations and Their Mechanisms Simulation and Verification of Tropical Deep Convective Clouds using Eddy-permitting Regional Atmospheric Models Improved Ocean State Estimation and Sensitivity Analysis Experiments for the Optimal Observing System, by using a 4D-VAR Ocean Data Assimilation System Yukio Masumoto Kozo Nakamura Shuhei Masuda RIGC, JAMSTEC RIGC, JAMSTEC RIGC, JAMSTEC 11 Global Elastic Response Simulation Seiji Tsuboi DrC, JAMSTEC Simulation Study on the Dynamics of the Mantle and Core in Earth-like Conditions Numerical Simulation of Seismic Wave Propagation and Strong Ground Motions in 3-D Heterogeneous Media Development of Advanced Simulation Tools for Solid Earth Sciences Numerical Simulations of the Dynamics of Volcanic Phenomena Yozo Hamano Takashi Furumura Mikito Furuichi Takehiro Koyaguchi 16 Space and Earth System Modeling Kanya Kusano Numerical Experiments with Multi-models for Paleoenvironmental Problems Model-observation Integration Study of the Middleatmosphere Dynamics Using a High-resolution Climate Model and the Antarctic PANSY radar Predictability Variation in Numerical Weather Prediction Ayako Abe Shingo Watanabe Takeshi Enomoto IFREE, JAMSTEC Center for Integrated Disaster Information Research, Interfaculty Initiative in Information Studies, The University of Tokyo/Earthquake Research Institute, The University of Tokyo IFREE, JAMSTEC Earthquake Research Institute, The University of Tokyo Space and Earth System Modeling Laboratory Unit, Laboratory for Earth Systems Science, JAMSTEC Atmosphere and Ocean Research Institute, The University of Tokyo RIGC, JAMSTEC Disaster Prevention Research Institute, Kyoto University 2

12 Project Name Name of Project Representative Professional Affiliation of Project Representative 20 Research on the climate change mechanisms in the Arctic regions based on improvements of the coldregion processes and validations of the Arctic climate reproducibility using global climate models Yoshiki Komuro RIGC, JAMSTEC Epoch-making Simulation (12 projects) Project Name Name of Project Representative Professional Affiliation of Project Representative Large-scale Simulation on the Properties of Carbonnanotube Large-scale Simulation for a Terahertz Resonance Superconductors Device Direct Numerical Simulations of Fundamental Turbulent Flows with the World's Largest Number of Grid-points and Application to Modeling of Engineering Turbulent Flows A Large-scale Post-genome Analysis using Self- Organizing Map for All Genome and Protein Sequences Development of a Fluid Simulation Approach by Massively Parallel Bits-operations with a New Viscosity Control Method Satoshi Nakamura Mikio Iizuka Takashi Ishihara Toshimichi Ikemura Hiroshi Matsuoka 26 Development of Adaptive High Accuracy Libraries Hidehiko Hasegawa Theoretical Study of Drug Resistance Mechanism Based on the Fragment Molecular Orbital Method Research Project of Biomedical Unsteady Fluidstructure Interaction Analysis for Cerebral Aneurysm and Other Organs Development of the Next-generation Computational Fracture Mechanics Simulator for Constructing Safe and Sustainable Society Precise Calculations for Few-Body Atomic Systems using the Gaussian Expansion Method and Applications to Cold Atoms Strongly Nonlinear Response Analyses on Building Structures with Large-scaled and Super-detailed Numerical Modeling Development of MHD heat transfer database at design conditions of advanced blanket system in fusion reactor Shigenori Tanaka Tadashi Tanuma Ryuji Shioya Emiko Hiyama Yoichi Mukai JAMSTEC : Japan Agency for Marine-Earth Science and Technology IFREE : Institute for Research on Earth Evolution ESC: The Earth Simulator Center RIGC : Research Institute for Global Change DrC : Data Research Center for Marine-Earth Sciences Research Organization for Information Science & Technology Research Organization for Information Science & Technology Graduate School of Engineering, Nagoya University Nagahama Institute of Bio-Science and Technology Research Institute of Electrical Communication, Tohoku University Faculty of Library, Information and Media Science, University of Tsukuba Graduate School of System Informatics, Kobe University Joint Program Center, Teikyo University Faculty of Information Sciences and Arts, Toyo University RIKEN Nishina Center for Accelerator- Based Science Graduate School of Engineering, Kobe University Yoshinobu Yamamoto University of Yamanashi 3

13 3. Collaboration Projects Collaboration Projects in FY2012 Institut Français de Recherche pour l Exploitation de la Mer (IFREMER), Département d Océanographie Physique et Spatiale, France Ernest Orlando Lawrence Berkeley National Laboratory, University of California (LBNL), USA Korean Ocean Research & Development Institute (KORDI), Korea The National Oceanography Centre, Southampton (NOCS), UK The large-scale numerical simulation of the weather/oceanographic phenomena for international maritime transportation : Kobe University Research and development for MSSG calculation performance optimization in the next-generation supercomputer system : RIKEN Collaborative research on the sophistication of the computational simulation software toward constructing the platform for the leading industrial research and development : Institute of Industrial Science, the University of Tokyo 4. System Configuration of the Earth Simulator The Earth Simulator - New Earth Simulator System of Ultra High-speed Vector Parallel Super Computer The Earth Simulator is the upgraded system of the previous Earth Simulator, which has significantly contributed to the development of a simulation culture in the area of earth science and related technical fields, and introduces new features to bring accurate and high-speed analysis and projections of global-scale environmental phenomena. The ES is also used to produce numerical simulations for advanced research fields that are beyond the scope of other computing systems. 4

14 5

15

16 Earth Simulator Research P roj ects

17

18 C h apter 1 Earth Science

19

20 Chapter 1 Earth Science U nderstanding Roles of O ceanic F ine Structures in C limate and I ts Variab ility I V Project Representative Wataru Ohfuchi Earth Simulator enter, apan gency for Marine-Earth Science and Technology Authors Hideharu Sasaki *1, unmei Taguchi *1, Akira Kuwano-Yoshida *1, Hidenori Aiki *2, Yoshikazu Sasai *2, Mayumi K. Yoshioka *3 and Wataru Ohfuchi *1 * 1 Earth Simulator enter, apan gency for Marine-Earth Science and Technology 2 esearch Institute for Global hange, apan gency for Marine-Earth Science and Technology * * 3 enter for tmospheric and ceanic Studies, Graduate School of Science, Tohoku niversity We have been investigating roles of oceanic fine structures in climate and its variability by using high-resolution, primitive equation based, global atmosphere, ocean and ocean-atmosphere coupled models, and a regional non-hydrostatic ocean-atmosphere coupled model. In this report, we present the following four topics, for all of which the oceanic fine structures play crucial roles. 1) Termination of aiu rainband, 2) Decadal variability of oceanic heat content in the orth acific, 3) Seasonality of submesoscales around the Kuroshio Extension, and 4) Momentum transfer from wind to surface waves and ocean currents in a coupled atmosphere ocean surface-wave model. Keywords oceanic fine structures, air-sea interaction, aiu rainband, oceanic heat content, oceanic submesoscales, surface wave 1. I ntroduction We have been leading researches on roles of oceanic fine scale structures in climate and its variations using atmosphere, ocean, and atmosphere-ocean coupled simulations on the Earth Simulator. Local air-sea interactions associated with oceanic fine structures such as fronts and eddies, which have significant potential impacts to modify local atmospheric and oceanic conditions and thus regional and larger scale climate systems, are observed in the world ocean. Oceanic fronts along the Kuroshio and Gulf Stream affect not only near-surface atmosphere but also entire troposphere, suggesting active roles of the mid-latitude ocean in the weather and climate. Furthermore, contributions of oceanic submesoscales smaller than mesoscale structures to large-scale oceanic field and oceanic ecosystems are implied in resent high-resolution satellite image and oceanic simulations. Oceanic surface wave and its dissipation should be considered to estimate surface momentum flux, which could reproduce realistic air-sea interactions with high surface wave. We highlight in this report our recent progress in researches on the roles of oceanic fine structures in climate and its variability using simulations. In Section 2, influence of the Kuroshio Extension to climatological termination of Baiu rainband is presented. new dynamics of orth acific oceanic heat content variability, in which oceanic frontal variability plays a central role, is proposed in Section 3. Seasonality of submesoscales around the Kuroshio Extension is reported in Section 4. Development of a coupled atmosphere ocean surfacewave model with an explicit treatment of momentum transfer from wind to surface waves and then to ocean currents under a tropical cyclone condition is reported in Section B aiu rainb and termination in AF ES and C F ES Baiu is a summer rainband stretching from eastern China through apan towards the orthwest acific. The climatological termination of the Baiu rainband is investigated using a standalone atmospheric general circulation model (AFES) forced with observed sea surface temperature (SST) and a coupled G M ( FES) with the same atmospheric component. The aiu rainband over the North Pacific abruptly shifts northward and weakens substantially in early July in AEFS (Fig. 1c) while it persists around 0 through summer in FES (Fig. 1b). The mid-troposphere westerly jet and its thermal advection explain this meridional position of the simulated Baiu, but ocean surface evaporation modulates the precipitation intensity. In AFES, deep convection in the subtropical Northwest Pacific sets in prematurely, displacing the westerly jet northward over cold ocean surface earlier than in observations (Fig. 2e). The suppressed surface evaporation over the cold ocean suppresses precipitation despite that mid-tropospheric warm advection and vertically integrated moisture convergence are similar to those before the westerly jet s northward shift (Fig. 2f). As 11

21 Annual Report of the Earth Simulator Center April March 2013 Fig. 1 Daily climatology of precipitation (color, mm day -1 ) and SST (contour interval is 2 with 10 and 20 thickened) averaged between 140 E and 170 E with 5-day running mean of (a) JRA25, (b) CFES and (c) AFES from 1 May to 30 Sep. Fig. 2 (left) 10-day mean between 21 July and 30 July of horizontal temperature advection (color, 10-6 K s -1 ) and vertical p-velocity with negative sign (contour interval is Pa s -1 with 0 and Pa s -1 thickened) at 500 hpa. (right) surface evaporation (color, mm day -1 ), SST (black contour, contour interval is 2 with 10 and 20 thickened), moisture flux at 925 h a (vector over 20 kg kg -1 m s -1 ) and vertically integrated moisture convergence (3, 5,, and 9 mm day -1 are plotted with blue contours). (a) (b) JRA25, (c) (d) CFES, and (e) (f) AFES. 12

22 Chapter 1 Earth Science a result, Baiu abruptly weakens after the northward shift in FES. In FES, cold SST biases in the subtropics inhibit deep convection, delaying the poleward excursion of the westerly jet. As a result, the updraft induced by the strong westerly jet and aiu both persist over the orthwest acific through summer in the FES (Fig. 2c). The results indicate that the westerly jet as well as ocean evaporation underneath is important for the Baiu rainband, suggesting a role of ocean in this important climate phenomenon. 3. Dynamics of N orth P acific oceanic h eat content variab ility on decadal time-scale pper ocean heat content ( ) is at the heart of the natural climate variability on interannual-to-decadal time-scales, providing climate memory and the source of decadal prediction skill. In the mid-latitude orth acific cean, signals are often found to propagate eastward as opposed to the frequentlyobserved westward propagation of sea surface height, a variable similar to OHC representing the ocean subsurface state. We investigate this dichotomy using a 150-year CFES control integration (Taguchi and Schneider ). Simulated signals are distinguished in terms of two processes that can support eastward propagation: higher baroclinic Rossby wave (RW) modes that are associated with density perturbation, and spiciness anomalies due to density compensated temperature and salinity anomalies. Our analysis suggests a unique role of the Kuroshio/Oyashio Extension (KOE) region as an origin of the spiciness and higher mode RW signals and we hypothesize a new mechanism for North Pacific decadal variability that links the westward- and eastward-propagating anomalies, as summarized in Fig. 3. First, wind-forced, westwardpropagating equivalent barotropic RW (Process 1. in Fig. 3) causes meridional shift of the subarctic front in the KOE region (2. in Fig. 3). The associated anomalous circulation crosses mean temperature and salinity gradients and thereby generates spiciness anomalies (3. in Fig. 3). These anomalies are then advected eastward by the mean currents (4. in Fig. 3), while the associated surface temperature anomalies are damped by air-sea heat exchange. Furthermore, the accompanying surface buoyancy flux generates higher baroclinic, eastward propagating W (5. in Fig. 3). The result suggests that the large OHC variability in the western boundary currents and their extensions is associated with the spiciness gradients and axial variability of oceanic fronts. The hypothesis derived from the FES simulation will be verified by future studies using rgo profiling float data, ocean reanalysis products, and other coupled CGCMs. 4. Seasonal variations of sub mesoscales around th e Kurosh io Ext ention ecent observations such as satellite observed SST and ocean color capture not only oceanic mesoscale phenomena (~ 100 km) of fronts and mesoscale eddies but also submesoscales ( 10 km) such as smaller eddies and elongated thin filaments. Recent studies of idealized high-resolution simulation also suggest that submesoscales induce strong vertical motions with fine horizontal scale and influence to large-scale and ecological fields. However, their spatial distribution and temporal variations in the basin scale ocean have not been revealed due to lack of observations. We have investigated submesoscales around the Kuroshio Extension and their seasonal variations in a high-resolution orth acific simulation using FES ( G M for the Earth Simulator) at 1/30 horizontal resolution 2. In the late winter, submesoscales are ubiquitous around the Kuroshio Extension (Fig. 4a). Vertical motions are strong in the mixed layer, which is deep in the season, and their horizontal scales are as fine as horizontal submesoscales. owever, the mesoscales larger than submesoscales are dominant in late summer (Fig. Fig. 3 Schematics of generation and propagation mechanism of decadal-scale ocean heat content ( ) anomalies in the orth acific. (a) mean (contours) and standard deviation (color shading) of annual mean sea surface height anomalies based on a 150-year long CFES control integration. Orange and red arrows schematically indicate westward-propagating, equivalent barotropic Rossby waves and the resultant latitudinal displacement of gyre boundary, respectively. (b) as in (a) but for OHC anomalies as measured with the temperature anomalies averaged over upper 400-m depth. Orange and red arrows schematically indicate eastward-propagating spiciness anomalies and higher barolcinic mode Rossby waves, respectively. 13

23 Annual Report of the Earth Simulator Center April March 2013 Fig. 4 Daily mean surface relative vorticity (10-5 s -1 ) on (a) March 15 and (b) September 15, 2002 in the orth acific FES at 1/30 resolution. b) when the mixed layer depth is shallow. The vertical motions with mesoscales mostly appear below the mixed layer. These results show seasonality of submesoscale activities around the Kuroshio Extension. Future studies are necessary to examine what mechanisms induce the ubiquitous submesoscales in late winter. We are also conducting this simulation with a simple ZD type biological model to see influence of submesoscales to oceanic ecosystem and its seasonal variability, which are our future studies. northeastward. The eye of surface waves (color) lags behind the eye of wind speed (contour), indicating that the regions of significant wind stress to ocean currents (i.e. the sum of the skin stress and the dissipation-induced stress) is somewhat different from the regions of high wind speed. 5. T ransfer of momentum from wind to surface waves and ocean currents in a coupled atmosph ere ocean surface-wave model The traditional bulk formula for wind stress on ocean circulation models is based on only 10 m wind speed, and is directed downwind. However under high wave conditions, the drag coefficient for wind stress might be better parameterized using quantities associated with surface gravity waves, such as significant wave height, wave age, and the direction of waves. Previous studies suggest that the net momentum flux from air (i.e. wind) to water (i.e. ocean current and surface waves) is given by the sum of the skin stress and the wave stress associated with the generation of surface waves, while the net momentum to ocean current is given by the sum of the skin stress and the dissipation-induced stress associated with the breaking of surface waves (Fig. 5). We have developed a coupled atmosphere ocean surface-wave model based on CReSS (Cloud Resolving Storm Simulator), NHOES (NonHydrostatic Ocean model for ES), and the surface-wave model of Donelan et al. (2012) 3. This coupled model adopts the dissipation (rather than the roughness) approach for estimating the momentum flux to ocean currents. byproduct of this approach is the availability of the dissipation rate of surface wave energy which is then used as the source term of the T E e uation for the oceanic mixed layer. We have investigated the impact of these effects on the hindcast simulation of tropical cyclones. Figure 6 shows a snapshot of a tropical cyclone translating Fig. 5 Schematic of momentum transfer from wind to surface waves and then to ocean currents ( iki and Greatbatch, 2013 ). Fig. 6 A snapshot of surface-wave significant height (m, color) and wind speed (m/s, contour) in a hindcast simulation of a tropical cyclone. 14

24 Chapter 1 Earth Science 6. C onclusion We have briefly reported research activities to investigate roles of oceanic fine structures in climate and its variability by using high-resolution, primitive equation based, global atmosphere, ocean and coupled models, and a regional nonhydrostatic ocean-atmosphere coupled model. In this fiscal year, the roles of oceanic mesoscales (fronts and eddies) as well as oceanic submesoscales and surface waves are investigated. We will study more on interactions between small and large scales and their influence on climate and its variability. References 1. Taguchi and. Schneider, rigin of Decadal-scale, Eastward-propagating Heat Content Anomalies in the North acific, submitted to ournal of limate. 2. Sasaki and. lein, SS Wavenumber Spectra in the orth acific from a igh- esolution ealistic Simulation, Journal of Physical Oceanography, 42, , M.. Donelan et al., Modeling waves and wind stress, Journal of Geophysical Research Oceans, 117, C00J23, doi / , iki and.. Greatbatch, new expression for the form stress term in the vertically Lagrangian mean framework for the effect of surface waves on the upper ocean circulation, ournal of hysical ceanography, (in press) 15

25 Annual Report of the Earth Simulator Center April March 2013 プロジェクト 責 任 者 大 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター 著 者 木 * 1, 文 明 * 1, * 1, 木 * 2, 一 * 2, 岡 * 3, 大 * 1 * 1 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター * 2 海 洋 研 究 開 発 機 構 地 球 環 境 変 * 3 大 大 研 究 大 海 洋 変 観 研 究 センター 海 洋 の 前 な の 空 間 的 に さいス ールの 現 が の 成 その 変 に す を 球 または の 大 海 洋 結 モデルを 用 いて 研 究 を 行 ている この 報 告 では 次 の つの 研 究 成 果 を り 上 た 球 大 モデルと 球 大 海 洋 結 モデルを 用 い 1) の と 球 大 海 洋 結 モデルを 用 い 2) 太 洋 の の 年 変 カニ ムを 明 らかにした また 高 解 度 太 洋 海 洋 モデルを 用 い 3) 流 の 海 洋 ブ ソ ス ール 現 の 変 を 明 らかにした さらに 4) のような 強 下 では の 生 成 と を して か ら 海 流 へ が されるという 論 をもとに 的 な 性 のとれた 大 海 洋 結 モデルを 開 発 した キーワード: 海 洋 の 構, 大 海 洋 作 用,, 海 洋, 海 洋 の ブ ソス ール 現, 16

26 Chapter 1 Earth Science Adaptation O riented Simulations for C limate Variab ility Project Representative Keiko Takahashi Earth Simulator Center, Japan Agency for Marine-Earth Science and Technology Authors Keiko Takahashi *1, yo hnishi *1, Yuya Baba *1, Shinichiro Kida *1, Keigo Matsuda *1, Li-Feng Lu *1, Youngjin Choi *1, Koji Goto *2 and Hiromitsu Fuchigami *3 * 1 Earth Simulator Center, Japan Agency for Marine-Earth Science and Technology * 2 NEC Corporation 3 E Informatec Systems TD * A coupled atmosphere-ocean-land model MSSG has been developed in the Earth Simulator Center, which is designed to model multi-scale interactions among the atmosphere, the ocean and the coupled system. Aiming to seamless simulation, cloud microphysics from the both view points of the accuracy of advection computation and high computational performance and oceanic part downscaling were improved and those results are summarized. In addition, three dimensional radiation scheme was developed and its impact was shown. The trial simulation for investigate regional climate variability was performed and its result was presented in this report. Keywords: Coupled atmosphere-ocean model, multi-scale, multi-physics, high performance computing, the Earth Simulator 1. I ntroduction Multi-Scale Simulator for the Geoenvironment (MSSG), which is a coupled atmosphere-ocean-land global circulation model, has been developed for seamless simulation based on multi-scale multi-physics modeling strategy in order to predict not only weather but climate variability. MSSG is optimized to be run on the Earth Simulator with high computational performance and it is designed to be available with flexibility for different space and time scales [1, 2, 3, 4]. In this report, summarizes a part of results of this project in FY2012 that focus on the following themes to execute seamless simulations with MSSG. In order to improve one of the main performance in seamless simulation, cloud micro-physics from the both view points of the accuracy of advection computation and high computational performance were improved. ceanic part downscaling is also important for seamless simulation with a coupled model MSSG. We developed the numerical model is based on a three dimensional particle random-walk model for improvement in the coastal region and validated improvement of physical performance in coastal region. Development of three dimensional radiation model regional climate model with MSSG for seamless simulation to consider an adaptation strategy for climate variability. 2. Advection sch eme improvement in cloud ph ysics process The weighted essentially non-oscillatory (WE ) scheme is applied to a cloud resolving model and is used for the cloud edge problem [5]. Validity is tested using three idealized experiments and the results are compared with those of recent height m height m (a) (c) Fig. 1 Simulation results comparison on cloud water distributions. (a)we5: (b) WE5-PUP: (c)ws5-pup: and (d)ws5-mup. 5m resolution was used [5]. height m height m (b) (d)

27 Annual Report of the Earth Simulator Center April March 2013 flux corrected transport (F T) schemes, i.e., D and M flux limiters. The WE scheme simulates the energy properties of cloud edges better than the FCT schemes do [5]. A one-dimensional advection condensation problem is first performed to investigate the advection scheme s ability to capture the cloud edge. The original WE scheme is found to be able to capture the cloud edge well. The WE scheme with D flux limiter avoids disadvantage that it produces a negative mixing ratio, and the scheme shows the smallest errors for all prognostic variables, especially for potential temperature. An ordinary D flux limiter cannot capture the potential temperature jump, since it was originally developed to avoid a negative mixing ratio. n the other hand, the accuracy of the M flux limiter is lower than that of the WE scheme because it has numerical diffusion at the cloud edge [5]. In a two-dimensional shallow cumulus convection experiment, the WE scheme was applied to the advection of mass and energy, accuracies for temperature and cloud water increase. Cross-sectional analysis on the shallow cumulus indicates that the cloud edge properties correspond to the overall temperature and cloud water trends, and the WE scheme simulates less evaporative cooling and cloud water evaporation. These facts indicate that the WE scheme also works well to capture the cloud edge compared to the FCT schemes (Fig. 1)[5]. Squall line experiment was performed to verify the effect of advection schemes in simulating deep convection. When the WE scheme is applied to advections of mass and energy, enhancements of both buoyancy and condensation, resulting in production of more water species, occur. It was found that the enhancements are caused by a larger and colder cold pool formed below the squall line, which is formed by the enhanced convective downdraught cloud mass flux due to the employed advection scheme [5]. order of It is a good start to assume the disturbance flow to be a Stokes flow. Wang et al.(2005) pointed out that the original superposition method ( rgsm, hereafter) does not satisfy the no-slip boundary conditions for multiple particles in the system. Ayala et al. (2007) developed an iterative superposition method (ItrSM, hereafter). ItrSM is more reliable but computationally expensive due to its iteration procedure. For example, it was reported that about 95% of the computational time was consumed for the ItrSM in a simulation for a system of 200,000 monodisperse particles in a turbulent flow on a 6 3 grid. The developed the binary-based superposition method (BiSM) proposes an intermediate method between rgsm and ItrSM in terms of both computational cost and reliability 6. In the binary-particle system shown in Fig. 2(a), the solution is directly obtained by BiSM and iteratively obtained by ItrSM. There is no error in BiSM compared to ItrSM. In a system containing three particles is shown in Fig. 2(b), BiSM ignores interactions via three or more particles 6. Figure 3 shows the collision efficiency E c between r 1 and r 2 particles in a stagnant flow. The solid line is from results which adopts ItrSM. The results from rgsm tend to produce larger values than ItrSM and BiSM. The consistency among ItrSM and 3. Development of h igh performance computational sch eme for droplet collision process: B ism ( th e b inary-b ased superposition meth od) While moving in a flow medium, a particle induces a flow disturbance in its neighborhood. The disturbance may intervene between particles for the so-called hydrodynamic interaction (HI). The particle Reynolds number based on the gravitational settling velocity for cloud droplets in the atmosphere is of the Fig. 3 ollision efficiency in stagnant flow. Fig. 2 (a) binary-particle system and (b) triplet-particle system. Fig. 4 Wall clock time versus number of cores for different number of particles and flow grids on TE, I E, and ES2. 18

28 Chapter 1 Earth Science Fig. 5 Simulation results of 134/137 Cs in seawater (Bq/L) at April 1st (left), May 1st (center), and June 1st (right), respectively [5]. ism confirms the reliability of ism for collision efficiency calculations in a stagnant flow 6. Under a typical dilute condition as in atmospheric clouds, the computational cost of BiSM is smaller by order of 10 than that of ItrSM, whereas an error of BiSM compared to ItrSM is insignificant. oupling of the cell-index method with ism can reduce the computational cost for paritcle interactions to (N pσ ), where N p is the total number of particles and 1 2, from 2 (N p ) (Fig. ) C oastal region downscaling improvement in ocean part of MSSG In the seamless simulation, oceanic part downscaling is also important scheme. For the improvement and validation of physical performance in coastal region, we developed the numerical model is based on a three dimensional particle random-walk model and a z-coordinate ocean general circulation model MSSG-. The simulation for radionuclide concentrations obtained from the density of particles per unit volume water was performed and validated. Experiments have been carried out for 137Cs for 4 months and the results show that coastal currents and meso-scale open oceanic eddies having large influence on the behavior of the radionuclides. The radionuclides in coastal currents remain along the coast where as the one in meso-scale open oceanic eddies rapidly escape to the interior of the acific along the uroshio extension. Depositions of radionuclides on sediments are mainly occurred during the first several months [7]. Figure 5 shows the distribution of 134/137 Cs in seawater. The initial peak is very well reproduced in the simulation results with our model MSSG-. too fast decrease in concentrations is again produced. During the first 2 weeks after the discharge, the great part of radionuclides flow into the Sendai Bay by the northward currents in the Sendai Bay. After 40 days, the radionuclides are distributed into the open oceans. In June 1st, the radionuclides discharge into the orth acific. The discharge route to the orth acific has two branches. The reason could maybe be related to the existence of the eddy which is observed and reproduced by the ocean model. The first one passes through the anti-cyclonic eddy while the other one follows the cyclonic path along the Kuroshio extension [7]. The most of sediment phase radionuclides deposit in the continental shelf regions (Fig. 6). This suggests that the radionuclides cannot penetrate into deep layers by advection and vertical mixing. The simulated radionuclides in bottom sediments may be caused by the rapid vertical mixing process in the ocean mixed layers [7]. 5. I mpact of th ree-dimensional radiation for urb an climate Factors of the heat island effect include three points of (1) increase in thermal capacity of the city with the change of the land use, (2) change of the emission heat transfer with the building and the convection heat transfer and (3) artificial exhaust heat by the use and the industrial activity of a car and the air conditioning. The impact of three-dimensional radiation which is described in (2) is analyzed. The three-dimensional radiation scheme for building-resolved simulation was improved which had been implemented in the MSSG model. The effect of three-dimensional radiative heat transfer on the urban thermal condition is investigated. Simulations started at 15:00 of August 5, 2005 and MSM data of the Meteorological Agency was used as a boundary condition. Figure 7 and 8 show the three dimensions spatial distribution of long wave emission flux toward the sky in 3D radiation (full S-to-S) and three-dimensional distribution of net longwave radiative flux into the surfaces, respectively. ecause sky factor in the alley put in the building is small, long wave emission to 19

29 Annual Report of the Earth Simulator Center April March 2013 Fig. 6 (a-f) adionuclide concentration in bottom sediment phase in simulations (a) March 30, (b) pril 1, (c) pril 29, (d) May 20, (e) une 1, and (f) June 29, (g-i) Monthly averages of observed radionuclide concentration in bottom sediments: (g) April, (h) May, and (i) June, 2011[7]. the sky is weak in Fig. 7. In Fig. 8, the net long wave emission flux shows negative value, on the other hand, there is the point indicating positive value in the alley put in the buildings. These results show that radiative cooling is restrained by buildings and that is one of the factor of heat storage in boundary layer. 6. MSSG as a regional climate model For the first step of seamless simulation to consider an adaptation strategy for climate variability, regional climate model was constructed using MSSG model to simulate urban climate over Kanto plain. Significant uncertainty in the atmospheric model is known to be cloud microphysics, and sensitivity of the urban climate to microphysics is examined. The simulation results with 4 km horizontal resolution show that two-moment microphysics improves trends of intense precipitation compared to one-moment scheme (Fig. 9). Further sensitivity study for horizontal resolution which was used with 4 km in these trial simulations is required for intense rainfall with inaccurate. 20

30 Chapter 1 Earth Science Fig. 7 Three-dimensional distribution of longwave radiative flux [W/m 2 ] toward the sky. Fig. 8 Three-dimensional distribution of net longwave radiative flux [W/m 2 ] into the surfaces. Fig. 9 Fre uency of precipitation mm/h form simulations in each summer during

31 Annual Report of the Earth Simulator Center April March F uture work In this report, we introduced improvements of physical performance of MSSG to performing seamless simulation. In near future, we are planning to validate physical performance of realistic multi-scale multi-physics phenomena such as M and El ino and Indian cean Dipole by longer integration with further high resolution. In the simulations, climate impact in urban area due to those climate variability will be investigated. References: [1] Keiko Takahashi, et al., World-highest Resolution Global Atmospheric Model and Its Performance on the Earth Simulator, roceeding of S 11 State of the ractice eports, Doi / , [2] Keiko Takahashi, et al., Non-hydrostatic Atmospheric G M Development and its computational performance, workshops/2004/ high_ performance_ computing-11 th / presentations.html, [3] Keiko Takahashi, et al., Non-hydrostatic atmospheric GCM development and its computational performance, Use of High Performance computing in meteorology, Walter Z wieflhofer and George Mozdzynski Eds., World Scientific, pp , [4] Yuya Baba and Keiko Takahashi, Large-eddy simulation of convective boundary layer with density stratification, Journal of the Meteorological Society of Japan, Vol.89, pp , [5] Yuya Baba and Keiko Takahashi, Weighted essentially nonoscillatory scheme for cloud edge problem, Quarterly Journal of the Royal Meteorological Society, 2012, doi: /qj yo nishi, eiko Takahashi, and.. assilicos, An efficient parallel simulation of interacting inertial particles in homogeneous isotropic turbulence, Journal of Computational Physics, Vol. 242, pp , [7] Youngjin Choi, Shinichiro Kida, and KeikoTakahashi, The role of oceanic circulation and mixing on the dispersion of radionuclides released from the Fukushima Daiichi uclear Power-Plant accident, Biogeosciences, 2013 (accepted). 22

32 Chapter 1 Earth Science プロジェクト 責 任 者 高 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター 著 者 高 * 1, 大 * 1, * 1, * 1, 木 一 郎 * 1,Li-Feng Lu* 1, Youngjin Choi* 1, * 2, 上 * 3 * 1 海 洋 研 究 開 発 機 構 地 球 シミュレータセンター * 2 NEC 会 社 * 3 NEC インフ マティックシステム 会 社 A coupled atmosphere-ocean-land model MSSG has been developed in the Earth Simulator Center, which is designed to model multi-scale interactions among the atmosphere, the ocean and the coupled system. Aiming to seamless simulation, cloud microphysics from the both view points of the accuracy of advection computation and high computational performance and oceanic part downscaling were improved and those results are summarized. In addition, three dimensional radiation scheme was developed and its impact was shown. The trial simulation for investigate regional climate variability was performed and its result was presented in this report. 変 現 による の への を 定 するためのシームレスシミュレーションを 可 能 とするという 最 目 のために MSSG-A( 大 大 環 モデルコード) MSSG- ( 海 洋 大 環 モデルコード) MSSG( 大 海 洋 結 モデルコード)にそれ れに 対 して にも 目 されている 数 の の 成 モデルの モデル 開 発 に 力 した MSSG-A においては モデルの 的 性 能 を する モデルにお る 高 度 の 計 算 法 を 開 発 し WEN ス キームで の 力 的 ル ーが 数 なく 流 され その 果 は ブグリッドス ール よりも 大 きいことを 示 し 高 度 流 スキームの 性 を 明 らかにした また の 法 では 膨 大 な 計 算 機 が なため 視 され てきた 流 体 を した 間 作 用 (Hydrodynamic Interaction, HI)を よく できる 法 BiSM(the binary-based superposition method)を 開 発 した また の イ による 強 度 の 異 を することにより 目 視 観 と の 明 を 再 現 した 結 果 の が 対 流 のス ール でなく より さなス ールでも 在 しており 実 の と の 度 変 が 再 現 できることを 示 した( 1) MSSG- においては Noh-Kim スキームの と 行 い が 強 く 表 が 数 的 に 定 である 問 題 を 解 し な 期 分 を 可 能 とした 加 えて 的 にも 数 が ない 海 と 海 の 間 でおきる 環 モデルへ し ウンス ーリングシミュレーションを 実 施 し での を 再 現 することに 成 した また マルチプロセス モデルとして 川 海 洋 の 作 用 入 した 地 上 流 モデルを 開 発 し データから 川 流 出 をシミュレーショ ン 可 能 にした( 2) と 変 のマルチス ール 関 性 を 明 らかにすることを 目 的 に ま の 特 性 を 再 現 する 3 次 元 計 算 スキームを 入 した 解 析 を 行 た 結 果 の を するためには 3 次 元 を す る 性 が 明 らかとな た さらに 関 の モデルの 構 を 行 い 最 もモデルの 実 性 が 強 いと えられ る スキームのイン クトを た 結 果 2 モー ントスキームが 強 度 分 でよりよい 結 果 を 示 すことが か た また 構 したモデルは 定 的 に 関 の および 地 表 度 を 再 現 できることが 分 か た キーワード: Coupled atmosphere-ocean model, MSSG, multi-scale, multi-physics, high performance computing, the Earth Simulator 23

33 Annual Report of the Earth Simulator Center April March ステレオグラム( 体 視 画 )で 示 した の 3 次 元 分 の はそれ れ 目 と 目 から 見 た 度 に 対 しており 法 による 体 視 として の 分 を することができる 26 回 数 流 体 シン ジウム スト CFD グラフィックスアワー ド 最 を 2 地 上 流 モデルによる 関 の の 地 表 の さ 分 24

34 Chapter 1 Earth Science Development of a H igh -Resolution C oupled C limate Model for Glob al War ming Studies Project Representative Akira Noda Research Institute for Global Change, Japan Agency for Marine-Earth Science and Technology Authors Akira Noda *1, Ayako Abe-Ouchi *2, 1, Megumi O. Chikamoto *1, 3, Yoshio Kawatani *1, Yoshiki Komuro *1, Masao Kurogi *1, Rumi Ohgaito *1, Fuyuki Saito *1, Kunio Takahashi *1, Kumiko Takata *1, 4 and Yukio Tanaka *1 * 1 Research Institute for Global Change, Japan Agency for Marine-Earth Science and Technology 2 Atmosphere and Ocean Research Institute, The University of Tokyo * * * 3 International acific esearch enter, niversity of awaii 4 Arctic Environment Research Center, National Institute of Polar Research The purpose of this project is to further develop physical models for global warming simulations, and to investigate mechanisms of changes in global environment as a successor of a previous ES joint project. We have obtained the following results this year. The T106L168 MIROC-AGCM without non-stationary gravity wave parameterization is constructed to investigate future changes of the quasi-biennial oscillation (QBO), semi-annual oscillation (SAO) and stratospheric sudden warming (SSW). These phenomena were well reproduced in the present climate simulation. The performance of an ocean model with a subgrid scale parameterization for multiple sea-ice thicknesses is investigated. This model reproduced ice-thickness distributions of the Arctic Ocean reasonably well. A numerical ice sheet model IcIES is extended to include ice-shelf/stream process. The implementation of the parallel ice-shelf model was succeeded, but there remained still many points to be improved. The non-parallel version of IcIES are applied for global warming experiment as well as Eemian experiment. The surface energy-water balance on land is evaluated by a wetness index (WI) using a global data (GPCC and NNRP) and the 20th century experiments. The mean global distribution of WI was reproduced better with MIROC5 than with CMIP3 GCMs. A simple scheme for the effect of surface soil moisture changes on soil-surface albedo is tested. As a result, surface albedo was increased in the dry regions, which could reduce the warm biases at the surface there. Keywords: Atmosphere-Ocean-Land coupled model, sea-ice thickness, stratospheric QBO, ice-sheet model 1. I ntroduction This project is a successor of one of the previous ES-joint projects named Development of a High-resolution Coupled Atmosphere-Ocean-Land General Circulation Model for Climate System Studies. The purpose of this project is to further develop physical models for global warming simulations, and to investigate mechanisms of changes in global environment. To achieve the purpose, we focus on the development of ice sheet model, permafrost model and sea ice model, improvement of subcomponent models for atmosphere, ocean and landsurface processes in the climate model MIROC, as well as sensitivity studies using climate models relevant to global warming and paleo-climate. 2. T h e Q B O, SAO and SSW simulated in th e MI RO C -AGC M The quasi-biennial oscillation (QBO), semi-annual oscillation (SAO) and stratospheric sudden warming (SSW) are reasonably simulated by the T106L168 MIROC-AGCM without parameterized nonstationary gravity wave forcing. The vertical resolution is set 500 m from the upper troposphere to the mesopause (~ 85 km altitude). The model is integrated for 50-years in the present climate condition. Figure 1 shows the time-height cross section of monthly mean and zonal mean zonal wind over the equator for 15-years. The amplitude and period of the simulated QBO are reasonable. The SAO is well simulated around 1 hpa. This model also successfully simulated the SSW events (not shown). It has been considered that the SSW occurs in the stratosphere. However, recent satellite observation indicated that the SSW also occurs in the mesosphere [1]. 25