A Physically Based Ground-level PM_2.5 Estimation Using Remote Sensing Technology

Over the past few years,regional air pollution has frequently occurred over Central and Eastern China with the development of urbanization and industrialization.Haze,mainly caused by the rising atmospheric fine particulate matter(PM2.5)concentration,not only leads to the decrease of atmospheric visibility,but also increases the mortality and morbidity of respiratory system diseases,which attracts much attention.However,the lack of enough ground-based PM2.5 sites and historical observations makes satellite remote sensing an effective means to monitor the concentration of PM2.5.The statistical models for PM2.5 retrieval can not be extended to other areas or historical period due to its dependence on ground-based observations and necessary auxiliary data.A physically based model is used in this paper to retrieve the ground-level PM2.5.This model is firstly evaluated by using the AERONET aerosol optical depth(AOD)data,which has higher accuracy compared to the satellite-derived AOD.Based on this physically based PM2.5 retrieval model,boundary layer height(BLH)and relative humidity(RH)reanalysis data from the European Centre for Medium-Range Weather Forcast(ECMWF)from 2015 to 2016 are used to make vertical correction and humidity correction for AOD products of AERONET sites,the ground-based PM2.5 data are used as reference data to validate retrieved PM2.5 results.The validation results of four AERONET sites in Beijing area are about 0.80 with relative coefficient(R).The best result appears at Beijing site,with R equal to 0.84 and Root Mean Squared Error(RMSE)is 33.67 ?g/m3,indicating that this physically based model has a good application potential.The advantage of this physically based model is that it has the ability to retrieve PM2.5 without ground-based PM2.5 observation and hence extendable in space and time.Using the C6 version of MOD04 data,and BLH,RH reanalysis data,ground-level PM2.5 concentrations are retrieved from 2007 to 2016 over Central and Eastern China based on this physically based model.Then,the ground-based PM2.5.observation data are used to validate the retrieved PM2.5 results.Result shows that the PM2.5 concentration retrieved by this physical method has high accuracy,which is applicable to the study of PM2.5 concentration over Central and Eastern China.The inter-annual variation and distribution of PM2.5 concentration over Central and Eastern of China are analyzed based on the space-bome retrieved PM2.5 concentration.Result shows that the annual averaged PM2.5 show a trend of gradually decreasing during 2007-2016,with the highest value in 2007,and lowest value in 2016.High values are found in the northwest plain in Shandong province,Huai River Plain,Poyang Lake Plain,and the Yangtze River Delta region(especially in areas along the Yangtze River).Very high values are found in Hebei Province.The low value is mainly distributed in the south-east China,and the lowest value appeared in Fujian province.