Remote sensing of aerosol optical depth from high spatial resolution images of HJ-1 constellation CCD sensors and application in the monitoring of environmental PM2.5

The high spatial and temporal resolution of HJ-1 (The China Environment and Disasters Monitoring Microsatellite Constellation) makes it suitable for environment monitoring. However, due to lack the mid-infrared wavelength near 2.1μm, it is difficult to use HJ-1/CCD images to retrieve aerosol optical depth (AOD) which is one of the important parameters in environmental monitoring. Many currently widely used AOD products are from MODIS and retrieved using dark object method, which has great restriction for its accuracy when applied in urban areas. In this paper, we present a new algorithm which makes use of HJ1-CCD blue and green bands to retrieve AOD without using other auxiliary data. This method is based on the assumption that aerosol optical properties change quickly with time but slowly with location, and based on "Multi-wavelength, Multi-sensor, Multi-day" (3M) to achieve the signal separation between the atmosphere and the earth. A LUT (look-up table) is constructed using 6S RT (Radiative Transfer) code, then the minimum of a cost function is used to determine the aerosol optical thickness. This paper describes the result of remote sensing of aerosol optical thickness based on this method in different regions (Beijing-Tianjin-Tangshan, Yangtze River Delta, Pearl River Delta, etc.) and different spatial resolutions (100m, 500m, 1000m, etc.) in China. The retrieved AOD results are then used to predict PM2.5 mass concentration in these regions. The AOD and PM2.5 results are in good agreement with the measured results from sun photometers and PM2.5 particulate samplers, respectively, which shows that the method has good accuracy and applicability.