A5080621827- earthquake and tectonic studies
- Seismic Imaging and Inversion Techniques
- Marine and coastal plant biology
- Meteorological Phenomena and Simulations
- Seismic Waves and Analysis
- Methane Hydrates and Related Phenomena
- Marine and fisheries research
- Seismology and Earthquake Studies
- Climate variability and models
- Earthquake and Tsunami Effects
- Model Reduction and Neural Networks
- Distributed and Parallel Computing Systems
- Fluid Dynamics and Turbulent Flows
- Food Industry and Aquatic Biology
- High-pressure geophysics and materials
- Geophysics and Sensor Technology
- Fluid Dynamics and Vibration Analysis
- Precipitation Measurement and Analysis
- Hydrology and Drought Analysis
- Marine animal studies overview
- Cryospheric studies and observations
- Infection Control and Ventilation
- Avian ecology and behavior
- Agriculture, Soil, Plant Science
- Icing and De-icing Technologies
RIKEN Center for Computational Science
2017-2024
Kobe University
2023-2024
Tokyo Metropolitan Agriculture and Forestry Research Center
2006-2024
Japan Agency for Marine-Earth Science and Technology
2014-2016
Institute of Agricultural Machinery
2008
Fisheries Agency
1998
We improved the tsunami simulation code JAGURS, which is a paralleled version of URSGA for large-scale, high-speed prediction in Nankai trough, Japan. optimized loop kernel velocity update and intergrid communication on three-dimensional torus network. Linear scaling was achieved up to full system capability K computer (82,944 nodes) strong test that used 100 billion finite-difference grid points. The measured performance 1.2 petaflops (11.5% peak speed). Intergrid three-nested-grid model...
This paper reports the implementation of high-performance computing using K supercomputer in Kobe, Japan, for large-scale/high-resolution flood simulation. Supercomputer was developed 2012 by RIKEN and Fujitsu ranked first list Top 500 sites 2011 during its development stage. A two-dimensional inundation simulation model based on a shallow water equation an existing numerical scheme parallelized with supercomputer. Osaka other cities along Yodo River were chosen as application area...
We present high-performance simulations of global seismic wave propagation with an unprecedented accuracy 1.2 s period for a realistic three-dimensional Earth model using the spectral element method on K computer. Our use total 665.2 billion grid points and resolve 1.8 trillion degrees freedom. To realize these large-scale computations, we optimize widely used community software code to efficiently address all hardware parallelization, especially thread-level parallelization solve bottleneck...
To validate the abundance of scenarios large earthquakes in Nankai Trough, we examined effects both lateral and vertical expansions source areas on maximum tsunami heights along Pacific coast Seto Inland Sea. The recently proposed Trough earthquake scenario (Mw = 9) has a slip 20 m near trough axis. However, predicted exceeded those obtained from historical records damage caused by 1707 Hoei event at Tosa Bay coastlines Kii Channel, owing to up-dip extension fault segments off Shikoku...
The fastest supercomputer in 2020, Fugaku, has not only achieved digital transformation of epidemiology allowing end-to-end, detailed quantitative modeling COVID-19 transmissions for the first time but also transformed behavior entire Japanese public through its analysis transmission risks multitudes societal situations entailing heavy risks. A novel aerosol simulation methodology was synthesized out a combination new CFD methods meeting industrial demands solver, CUBE (Jansson et al.,...
Future changes in large-scale climatology and perturbation may have different impacts on regional climate change. It is important to understand the of terms both thermodynamic dynamic changes. Although many studies investigated influence climate, significance still debated. The nonlinear effect these two also unknown. We propose a systematic procedure that extracts influences three factors: climatology, resulting effect. then demonstrate usefulness procedure, applying it future...
Regional climate projections inevitably inherit uncertainties from general circulation models (GCMs). We therefore propose a new approach for identifying the dominant uncertainties. This employs downscaling procedure by Adachi et al. to uncertainty problem using multiple GCM projections. The mean state of large‐scale atmospheric states and deviation this are two factors considered here, which provided GCM. These referred as climatology perturbation components, respectively. To demonstrate...
The robustness of an artificial neural network that performs model order reduction for flow field data is studied. trained with a large-scale distributed learning approach using up to 6,259 nodes the supercomputer Fugaku. Flow around two square cylinders varying distance between their centers investigated. and tested from numerical simulations. First, capability reproduce fields 2, 12, 24 modes investigated by comparing reconstructed simulated data. It shown, reconstructions based on 2...
In this study, we provide a perspective on dynamical downscaling that includes comprehensive view of multiple methods and strategy for achieving better assessment future regional climates. A climate simulation is generally driven by large-scale atmospheric state obtained global simulation. We conceptualize the based reconstruction combining decomposed components states, such as climatology perturbation, in different simulations. The conceptualization provides previous studies. propose...
This article presents a case study on the extension of parallel algorithms in tsunami and earthquake-cycle simulators for massively execution K computer. We use two target applications: tsunami-simulation program, “JAGURS,” an “RSGDX.” Our optimization strategy collective communication is to split Message Passing Interface (MPI) communicator perform multistage localized minimize frequency, transferred data size, network congestion. Moreover, severe load imbalances, we apply cyclic...
Abstract Gut contents and nitrogen isotope ratio ( δ 15 N) analyses were conducted on splendid alfonsino Beryx splendens collected in the vicinity of Kuroshio from May 2015 to November 2016 n = 708) identify their feeding habits Western North Pacific. Micronektonic fish, shrimp, squid dominant prey species, with Decapoda Acanthephyra quadrispinosa , Prehensilosergia prehensilis Cephalopoda Enoploteuthis chunii Chiroteuthis calyx Teleostei Sigmops gracilis Gonostoma elongatum . These...
Higher fidelity seafloor topography and crustal structure models have become available with accumulation of observation data. Previous studies shown that the consideration such high-fidelity produces significant effects, in some cases, on deformation results are used as inputs for tsunami analysis. However, it is difficult to apply model computations because large computational costs. In this paper, we propose a new computation method estimating computations, which based finite element...