反演同化技术在激光雷达中的应用研究

为增强模式预报准确性,提高激光雷达数据在化学模式中的应用水平,建立激光雷达同化系统,将激光雷达数据引入模式,利用WRF-Chem模拟2016年12月6日~10日一次污染过程,将激光雷达监测的气溶胶光学深度(AOD)反演成近地面PM2.5浓度,通过三维变分同化技术订正模式原始场,调整模式预报结果。实验对比发现,反演同化后水平和垂直方向初始场调整显著,PM2.5、PM10和AQI的预报准确率明显提升,同化后相关性提高了0.05~0.1,预报准确率提高了10%~20%,分级风险评分也不同程度升高,与观测结果接近,优于不同化的预报结果。表明激光雷达数据在模式同化方面表现良好,有广阔的应用空间。

Research on the Application of Inversion Assimilation Technology in Laser Radar

In order to enhance the accuracy of model prediction and improve the application level of laser radar data in chemical models,a laser radar assimilation system is established to introduce laser radar data into the model.In this paper,WRF-Chem model was used to simulate a pollution process during the period from 6 to 10 December 2016.The aerosol optical depth(AOD)monitored by the laser radar was inverted into PM2.5 concentration near the ground,and the initial field of the model was revised through three-dimensional variational assimilation technology to adjust the model prediction results.The result showed that the horizontal and vertical initial field changed obviously after assimilation,and the prediction accuracy of PM2.5,PM10 and AQI were greatly improved,the correlation was increased by about 0.05 to 0.1,the prediction accuracy was increased by about 10% to 20%,and the graded risk score also rise after assimilation,which was close to the observed results and better than that of non-assimilation.The experiment showed that the laser radar data perform well in model assimilation and have wide development space.