二维平流扩散模型下的集合卡尔曼滤波模拟同化研究

发布时间：2018-05-10 13:12

本文选题：数据同化 + 平流扩散模型　；参考：《中国环境监测》2016年06期

【摘要】：为提高环境数值预报水平,构建了一个针对污染物扩散的模拟数据同化系统。采用集合卡尔曼滤波方法对二维平流扩散模型的状态变量进行了实时校正,实现污染物浓度的实时模拟预报,完成了敏感性实验中集合数目变化、观测方差变化和同化窗口长度变化研究。比较考察观测点位置与污染源距离不同时的预报效果,探讨了优化条件下的同化策略,提出一种根据距离远近动态调节卡尔曼增益权重的方法。在集合数目较小时,可降低计算代价,得到优化的同化效果。
[Abstract]:In order to improve the environmental numerical prediction level, a simulated data assimilation system for pollutant diffusion was constructed. The state variables of two-dimensional advection diffusion model are corrected in real time by means of set Kalman filter, and the concentration of pollutants is predicted in real time, and the number of sets is changed in the sensitivity experiment. The variation of observation variance and the length of assimilation window are studied. In this paper, we compare the prediction results between the location of the observation point and the distance of the pollution source, discuss the assimilation strategy under the optimized conditions, and propose a method to adjust the Kalman gain weight dynamically according to the distance. When the number of sets is small, the computational cost can be reduced and the optimized assimilation effect can be obtained.
【作者单位】：西北师范大学物理与电子工程学院;
【基金】：国家自然科学基金(41461078) 兰州市科技计划项目(2015-3-34)
【分类号】：X51;X55;X830

【相似文献】