WRF模式计算效率的多因素影响机制及其优化方法研究  

Research on the Multi-Factor Influence Mechanism of Computational Efficiency of WRF Model and Its Optimisation Methods

作　　者：杜江付 李瑞娟 宁嘉泉 张俊兵 施晨晓 买合木提江·维吉坦 王露践 白磊 

机构地区：[1]海南大学生态学院,海南 海口 [2]海南省热带生态环境修复工程研究中心(海南大学),海南 海口 [3]山西省大气探测技术保障中心,山西 太原 [4]海南省气象信息中心,海南省南海气象防灾减灾重点实验室,海南 海口

出　　处：《计算机科学与应用》2024年第11期226-235,共10页Computer Science and Application

摘　　要：通过设置不同实验方案,运行WRF模型,分析结果文件来评估计算WRF所需的迭代时间受到的多种因素影响,如选取的网格大小,水平分辨率,垂直分辨率,处理配置,通过改变这些影响因素分析他们对WRF计算的时间和内存占用的影响。具体来说:较大的网格大小,较高的空间分辨率(水平方向与垂直方向)会显著增加计算所需的时间,而选取不同的处理配置只在较大的网格条件下才会有明显的影响。本文的研究为提高WRF模式的计算效率提供了理论依据和实践指导,对于类似大型并行应用的性能优化也具有一定的参考价值。By developing different experimental schemes and running the WRF model, this study analyzes the output files to evaluate the impact of various factors on the iteration time required for WRF computations, such as the selected grid size, horizontal resolution, vertical resolution, and processing configuration. By altering these factors, the study analyses their effects on computation time and memory usage. Specifically, larger grid sizes and higher spatial resolutions (both horizontal and vertical) significantly increase computation time, while different processing configurations only have a noticeable impact under conditions of larger grids. This research provides a theoretical basis and practical guidance for improving the computational efficiency of the WRF model and offers valuable insights for performance optimization in similar large-scale parallel applications.

关 键 词：WRF计算时间 水平分辨率 垂直分辨率 处理配置 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]