Improvement Of Cloud Detection Algorithm In High Pollution Area And Fine Simulation Of Its Porduct Trajectory Diffusion

With the rapid development of China's economy and urbanization,the problem of atmospheric environment is becoming more and more prominent.Pollutant monitoring using the satellite remote sensing technology has been widely used,but the high pollution area and cloud confused identification directly affects the spatial coverage and inversion of AOD products.At the same time,it is important to construct a highresolution regional wind field information,simulate and monitor the trajectory diffusion of AOD products with data assimilation technology,which have important application value for understanding the sources and sinks of pollutants.Therefore,this study uses the naive Bayesian cloud detection method to improve the identification of high pollution areas and cloud areas,which can effectively change the coverage of AOD inversion products.The high-resolution wind field information was built based on the WRF model and data assimilation technology.Then,the wind field information was used as input to achieve high-resolution air pollution diffusion trajectory simulation in high-pollution areas with the trajectory mode HYSPLIT,which can provide technical reference for monitoring and early warning of high-pollution weather.The main conclusions of this article include:(1)The Naive Bayesian cloud detection algorithm can effectively distinguish between high pollution areas and clouds,and improve the accuracy of cloud masks.In addition,it has been confirmed by VIIRS true color images for qualitative verification,and CALIPSO VFM data for quantitative verification in improved area.The results show that the improved cloud detection method is stable and effective,and can be extended to other satellite cloud mask algorithms such as MODIS.At the same time,the improved cloud mask products can effectively increase the coverage of AOD inversion results,especially in high pollution areas where people express more concerns.(2)In this study,the regional model was constructed by using the mesoscale model WRF,and conventional observation data were assimilated based on the GSI assimilation system,which effectively improved the accuracy of the analysis field.According to the analysis result of wind field data assimilation,the improving effect occurs on the near-surface and the 850 hPa wind field.Furtherly,the study analyzes the atmospheric circulation situation field.The main meteorological conditions of the pollution include the flat westerly wind in front of the 500 hPa high pressure ridge,the anticyclone at 850 hPa and the cold surface high pressure.In the end,it is found that the circulation situation field of the air pollution incidents in the study will constrain the horizontal and vertical airflow transport in the area,which is not beneficial for the pollution diffusion.(3)The research uses the HYSPLIT trajectory model,takes the AOD products before and after improvement and the wind field before and after assimilation as its input,and conducts a comparative experiment to compare the horizontal and vertical trajectories of AOD diffusion and the changes in concentration.It is found that AOD products mainly have a greater impact on the initial position at the beginning of the simulation,and the difference of wind field input becomes more major at a later time.The effect of concentration changes is more complicated than that of the diffusion trajectory,but the two influence stages change less obviously.