Estimating an accurate initial condition is very important for numerical weather prediction (NWP) because the weather system is known to be sensitive to the initial condition. Data assimilation plays an essential role in NWP to estimate the best initial condition using both simulation and observation data with their error statistics provided. Data Assimilation Research Team has been developing global and regional NWP systems known as the NICAM-LETKF and SCALE-LETKF, in which the ensemble Kalman filter, an advanced data assimilation method, is used. In this talk, I will present a brief introduction to data assimilation in NWP, and will convey some flavor of cutting-edge research such as precipitation data assimilation.

FDPS is a library for massively parallel particle simulations, which has been developed by our team. The current released version of FDPS can generate a parallel code that scales up to the K computer (Iwasawa et al. 2015, PASJ, 68, 54). However, it is not clear that FDPS generate a scalable code in the next generation of supercomputers. In order to identify potential problems of FDPS and to improve FDPS more, we need to analyze FDPS in a system larger than the K computer. In late January, we started the development of FDPS on Sunway Taihulight, which is the most powerful supercomputer in the world (as of Nov. 2016, 125.4 PFlops). In this talk, I'd like to introduce the current status of this project.

In the recently, the scale of supercomputer has been increasing remarkably. This trend will have been going from now. On the other hand, in network, multiple communication devices are installed and fat-tree, mesh/torus as network topology are adopted. There are several problems to be solved for adopting these. One is the efficient utilization of multiple communication devices. Another one is shortening communication delay by reducing the number of hops or contentions on a such network topology. We will introduce the optimization of communication for these problems.

Interaction between solar wind(supersonic plasma flow from the Sun) and intrinsic magnetic field of planets such as Earth forms plasma circulation system called "magnetosphere". A great number of studies have been done for the Earth magnetosphere using spacecraft observations, numerical simulations, and modelings. On the other hands, comparative studies of planetary magnetosphere such as Mercury or Jupiter are also important from a perspective of universal plasma physics. In this study, we try to discover the dynamics of Mercury's magnetosphere using magnetohydrodynamics(MHD) simulation, and trajectory tracings of test particles.

After flood disaster happened in Kinugawa river basin in 2015, citizens in Mukogawa river basin warried about torrential rain in the basin and they wanted to know what kind of damage they would suffer when it should be similar rain. The presentation explains the result of simulation with several introductions including Japanese river improvement plan and implementation, newer trial of computation science for dam utilization and contribution of science and technology to evacuation activity of citizens.

In Quantum Chemistry, usually we use the molecular orbitals to discretize the physical space. In that case, the dimension of the Hilbert space would increase combinatorially with the system size. To solve this problem, there are suggested variety of Quantum Chemistry algorithms so far. We are inventing new formalism that would use several BCS-like states to describe the molecular systems. With usage of explicit mathematical operations, we can describe the Quantum state effectively. We will introduce the recent developments with a comparison towards the already established algorithms.

I have low self-esteem. Actually the results of tests for measuring self-esteem show that my self-esteem is low level. In such circumstances, I have been interested in raising my self-esteem. On one occasion, I met with an Adlerian psychological literature named “The Courage of Being Disliked”, and I found three important keywords many times: proprioception, contributions of others and trusting of others, in it. Adler said that the contribution of others is an essential part in raising one’s worth. What is the “contribution of others”? I would like to introduce the Adlerian psychology and discuss the issue with computational scientists.

For the last several decades, Lagrangian finite element methods, which use deformed-mesh, have been the de facto standard for solid/structure simulations. This method can track the surface of solid with high accuracy and compute constitutive relation of solid with ease, since a geometric domain of the solid is spatially discretized using unstructured mesh which conforms to the deforming solid.
On the other hand, an Eulerian finite element method for solid/structure mechanics was proposed in the early 1990s. This method uses spatially fixed mesh and thus allows solid material to flow through the mesh. In this context, compared to Lagrangian finite element methods (Deformed-mesh methods), Eulerian finite element methods (Fixed-mesh methods) are attractive for computing large deformation problems, computing fracture problems, and generating computational mesh easily.
Utilizing the above advantages, Eulerian finite element methods have been applied for bullet impact analysis, solid-fluid interaction analysis, soft adhesive analysis, and soft resin analysis. However, the computational challenges have been their expensive computational cost, the diffusion of the solution due to the advection, and the modeling of contact interfaces. In my presentation, I will explain the past, present and future studies of fixed-mesh methods for solid/structure mechanics.

Sort-last visualization approach has become the defacto standard for massively parallel rendering environments. In this approach, the entire set of the images generated by the rendering nodes needs to be combined into a single final image, and for this purpose, parallel image compositing algorithms which try to minimize the idle time of the nodes have been widely used. As the number of nodes increases, the parallel rendering processes can have their workload reduced. However, for the parallel image compositing process, the workload increases proportionally with the increase in the number of rendering nodes, which is similar to the weak scaling case. In this talk, I will present the parallel image compositing library, named 234Compositor, primarily developed for the HIVE visualization application which works on the K computer massively parallel environment.

How should the rain forecast simulations be evaluated when we compare the simulations with the observation data? Is it possible to evaluate the simulations comprehensively by the shape and location of rainfall areas? The commonly used verification methods for the rain forecast have difficulties to evaluate the shape and location error of the rainfall areas. If the forecast has location errors even though the shape of rainfall areas is predicted perfectly, these methods evaluate the forecast lower. This evaluation does not match with our intuition.
Therefore, we propose a new verification method for the rain forecast using the shape and location of the rainfall areas. In this method, to mimic the human recognition, the close rainfall areas are combined, and the features of shape and location of the rainfall areas in the combined areas are computed. The proposed method can evaluate the shape and location error that cannot be evaluated by the frequently used methods. Additionally, by using the visualization with colors, this method can verify the rain forecast intuitively.