| Aerosols over urban areas not only affect the physical and mental health of urban residents,but also worsen urban traffic conditions.The spatial distribution of urban aerosols can be easily and quickly received by satellite remote sensing.When using the wide field viewing(WFV)camera of Gaofen-1(GF-1)satellite to monitor the aerosol over urban areas,two important problems need to be solved,one is to remove the contribution of urban complex surface and the other is to improve the composition of urban aerosol.In this dissertation,we use the surface spectrum data and MODIS surface reflectance products(MOD09A1)to analyze the reflectance of urban surface,assume that there is a linear relationship between red and blue bands to remove the surface contribution,use Aerosol Robotic Network(AERONET)data to improve the composition of urban aerosol,and study the retrieval urban aerosol from WFV data.It mainly includes the following four aspects:(1)The study of urban surface reflectance in red and blue band.Based on the spectrum data of vegetation and cement land in four ground experiments over urban areas of Beijing,it was found that there was an obvious linear relationship between the surface reflectance of typical land surface in red and blue bands;from the MODIS MOD09A1 products from 2013 to2017,four typical urban surface(water body,bare soil,building and vegetation)were extracted,and it was found that the yearly correlation between red and blue bands was higher than the seasonal correlation;four cities(Beijing,Tianjin,Shijiazhuang and Baoding)were selected to carry out pixel by pixel correlation analysis of surface reflectance in red and blue bands from2013 to 2017 with MOD09A1.The results showed that over more than 84% of the surface,the correlation coefficient of surface reflectance between red and blue bands was greater than 0.75.(2)The development of the retrieval algorithm of urban aerosol from the WFV data of GF-1 satellite.Using AERONET product and Optical Properties of Aerosols and Clouds(OPAC)database,the optical properties of mixing three basic aerosols(water-soluble,dust and soot)and ground observation were best fitting,then the urban aerosol model were improved,which was water-soluble: 80%,dust: 18% and soot: 2% respectively.After the preprocessing of WFV data of GF-1 satellite,based on the linear coefficient between red and blue band surface reflectance for every pixel over urban,the urban aerosol inversion algorithm was developed,and then the aerosol optical depth(AOD)in urban areas of Beijing,Tianjin,Shijiazhuang,and Baoding from July 2016 to September 2017 was inversed,and the inversion error caused by the difference of calibration coefficients,observation geometry and sensor spectral response was analyzed.(3)The comparison of our algorithm results with the results of deep blue(DB)algorithm,MODIS aerosol products,and ground observation.The AOD results of DB algorithm and MODIS products in Beijing,Tianjin,Shijiazhuang,Baoding from July 2016 to September 2017 were received,and compared with the AOD results of our algorithm in this dissertation.The comparison showed that our algorithm had a high correlation with DB's AOD and MODIS aerosol products,which was higher than 0.9.The AOD of four AERONET stations in Beijing urban areas during the same period were extracted,then the AOD was validated.It was shown that the correlation coefficient between our algorithm results and the AERONET observations was more than 0.907,which was better than DB results and MODIS aerosol products.(4)The application of our retrieved AOD to the PM2.5 monitoring in urban area.From the AOD from 2016 to 2017 retrieved by our algorithm,the meteorological data and PM2.5concentration of ground monitoring stations in the same period,the PM2.5 concentration in Beijing area was estimated by multiple linear regression mehtod,and the spatial distribution of PM2.5 concentration in Beijing urban area was discussed. |
PDF Full Text Request