Rsearch On Inversion Of PM_2.5 Mass Concentration In Xi’an Based On MODIS And Ground Monitoring Data

With the development of industry,transportation and urbanization in China,more pollutants are released into the atmosphere,causing a large amount of pollutants gathered in the atmosphere to form fog/haze pollution.At present,monitoring of haze pollution in China mainly relies on the conventional ground monitoring stations for observation.However,the number of monitoring station is small and sparsely distributed,which is limited in time and space,and can not meet the needs of regional and efficient monitoring of fog/haze pollution.The satellite remote sensing monitoring technology has obvious advantages in monitoring the continuity,spatial distribution and trend of the air pollution change because of its large-scale,continuous and dynamic characteristics.Satellite remote sensing technology can make up for some deficiencies of ground monitoring technology,but also for ground monitoring data to make a powerful supplement,which largely promotes the development of fog/haze monitoring technology.Therefore,it is of great practical significance to use satellite remote sensing technology to monitor and control the fog/haze pollution.In this paper,the atmospheric water vapor and aerosol data obtained by MODIS sensor and the main air pollutants and meteorological data acquired by ground monitoring stations are used to explore the main components of fog/haze pollution in Xi’an and their law of change with various factors.Based on the aerosol optical depth（AOD）retrieved by MODIS remote sensing,the relationship model between PM_2.5.5 mass concentration and AOD,atmospheric water vapor（PWV）and meteorological factors was studied in order to achieve the objective of retrieving PM_2.5.5 by remote sensing.The main research contents and the corresponding conclusions are as follows:1.AOD and PWV were extracted from the image data of MOD04_L2 and MOD05_L2by ENVI+IDL method.The daily variation trend of PWV and the average water vapor content in four seasons in Xi’an in 2016 were analyzed.It was found that PWV content had obvious seasonal changes.Among them,the average PWV content in summer is about 23.25mm,the average PWV content in spring and autumn is about 8mm,the average PWV content in winter is at least about 3.32mm,and the PWV content in summer is much higher than the other three quarters.The reason is mainly The western Pacific subtropical high affects the transport of atmospheric water vapor content and the blending of warm and cold air.2.Through analysis of the content of haze contaminants throughout the year and seasons and their changing characteristics,the PM2.5 mass concentration was selected as the evaluation index of haze pollution.Combined with the monitoring data of the ground monitoring station,the variation regularities of PM_2.5.5 mass concentration,meteorological factors,PWV content and AOD were studied.It was found that there was a certain correlation between PM_2.5.5 concentration and other factors in the daily value data comparison.The results showed that there was a positive correlation between PM_2.5.5 mass concentration and air temperature and relative humidity,negatively correlated with air pressure,wind speed and precipitation;PM_2.5.5 mass concentration had a positive correlation with PWV content;PM_2.5mass concentration positively correlated with AOD Relationship but weakly related.3.Since AOD was measured in the atmosphere and PM_2.5.5 mass concentration was measured near ground with a drying process,in order to improve the correlation between PM2.5 mass concentration and AOD,vertical correction and humidity correction were performed for the AOD and PM_2.5.5 mass concentrations.The corrected AOD and PM_2.5.5 mass concentrations were studied and found that the two have a good correlation.However,the relevance of the summer has been reduced,therefore,this paper will not make summer humidity correction,only a vertical correction.4.Regression modeling of the revised AOD and PM_2.5.5 mass concentrations was carried out by using three kinds of function types such as quadratic,exponential and power function.From the coefficient R² of the model,we can see that the power function model is the best in winter and the one quadratic function model in spring,summer and autumn are better than the other two function models.5.Taking into account the fog/haze change process is complicated and affected by a variety of factors,the"backward screening"method was used to model the corrected PM_2.5mass concentration,AOD,PWV content,meteorological factors and so on,and then verified and analyzed.From the model’s coefficient of determination R² and the estimated relative error of PM_2.5.5 mass concentration,the PM_2.5.5 remote sensing estimation model based on multiple factors is better than the remote sensing estimation model based on AOD.