High Performance Computing Aspects of Numerical Solvers for Flow and Transport in Climate and Environment Applications

Multiscale problems in Geosciences, in particular simulations of flow and transport in ocean, atmosphere, and subsurface require massive computational resources and belong to the shortlist of potential target applications for Exascale computing. However, in order to be able to efficiently use the future computational infrastructure, the numerical methods, the programming techniques, and the code development workflows must be first made capable of running efficiently on various hardware architectures (CPUs, GPUs, FPGAs, etc.) on one hand and, on the other, provide user-friendly model development interfaces in the form of domain specific languages (DSLs) or code generation frameworks. The goal of this topical collection is to document the current state-of-the-art in various application areas, numerical mathematics, and high performance computing and to discuss the recent trends and future directions for the numerical model development methodology and improvements in computational efficiency. The topics include, in particular, numerical and algorithmic approaches for high performance computing and emerging hardware architectures, parallelization strategies, hybrid computing architectures, DSLs, etc.