Comparison of adaptive mesh refinement techniques for numerical weather prediction

Daniel S. Abdi,Ann Almgren,Francis X. Giraldo,Isidora Jankov

2024-04-25

Abstract:This paper examines the application of adaptive mesh refinement (AMR) in the field of numerical weather prediction (NWP). We implement and assess two distinct AMR approaches and evaluate their performance through standard NWP benchmarks. In both cases, we solve the fully compressible Euler equations, fundamental to many non-hydrostatic weather models.

Numerical Analysis

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the effectiveness and superiority of applying Adaptive Mesh Refinement (AMR) techniques in Numerical Weather Prediction (NWP). Specifically, it explores the performance of two different AMR methods in NWP and evaluates these methods through standard NWP benchmark tests. These two methods are respectively the method that combines tree - structured cell refinement with spatial discretization of the high - order discontinuous Galerkin method, and the method that combines level - based AMR with the finite - difference method.

The paper mainly focuses on the following points:

1. **Multi - scale problems**: Traditional uniform grids are difficult to handle the multi - scale characteristics in atmospheric phenomena, especially when it is necessary to capture small - scale features, such as tropical cyclones and convective storms. AMR techniques can dynamically adjust the grid resolution and optimize the allocation of computing resources.

2. **Mass and energy conservation**: Study the influence of different AMR methods on mass and energy conservation to ensure the physical rationality of the model simulation results.

3. **Scalability and performance portability**: Examine the performance of AMR methods in large - scale parallel computing environments and their cross - platform applicability, which is crucial for practical applications.

4. **Solution transfer strategy**: Propose a simple and memory - efficient solution transfer strategy applicable to tree - structured AMR methods to ensure flow conservation in box - shaped and spherical regions.

5. **Possibility of replacing existing methods**: The grid refinement method commonly used in the current NWP field is the static nested grid. This method is relatively simple but not flexible enough. This study hopes to promote the application of more advanced dynamic AMR techniques in NWP through the research on the AMR framework (such as AMReX), so as to better track locally important features (such as tropical cyclones) and achieve higher scalability and performance portability.

In conclusion, this paper aims to explore the application potential of AMR techniques in NWP and provide new ideas and technical support for improving the accuracy of weather forecasts.