Implementation of a GPU-Accelerated Lagrangian Particle Dispersion Model for Atmospheric Transport of Radioactive Nuclides

摘要: Large-scale atmospheric dispersion modeling for emergency response to nuclear accidents requires both high computational efficiency and numerical reliability. A GPU-oriented Lagrangian particle dispersion modeling was developed within the FLEXPART framework to address these demands. Core transport processes—including advection, turbulent diffusion, convective mixing, and dry and wet deposition—were restructured for GPU parallel execution, achieving efficient operation on modern accelerator architectures. Further incorporation of fast arithmetic operators and multi-level parallelization strategies substantially improved overall computational performance while preserving physical accuracy. Additional MPI-based parallel meteorological data decoupling and preprocessing tool has been developed, which alleviates data-handling bottlenecks. While multi-GPU execution and a load-balancing strategy enable efficient scaling in heterogeneous computing environments. Based on the European Tracer Experiment’s first trial (ETEX-I), the computational accuracy and acceleration performance of the developed GPU program were rigorously evaluated. The results demonstrate that the program achieves simulation accuracy comparable to the benchmark reference while exhibiting modest improvements in the representation of turbulent mixing under certain conditions. Performance assessments indicate an overall speedup of approximately 40.45 on a single-GPU platform, which can be further increased to about 52.05 in high-performance application scenarios where meteorological background fields are reusable. Moreover, multi-GPU experiments reveal favorable parallel scalability across configurations ranging from one to four GPUs, and confirm that the proposed load-balancing strategy effectively enhances computational efficiency in heterogeneous GPU environments.