Aerosol Retrievals Over Bright Urban Surfaces Using Landsat 8 Images

Abstract

Aerosols, suspended particles in the atmosphere, play an important role in the formation of cloud nuclei, climate change, poor air quality and degradation of atmospheric visibility, and public health and diseases. For a full understanding of aerosols effects on urban air quality and public health, retrieval of aerosol optical properties such as aerosol optical depth (AOD) over bright urban surfaces is required. In this study, Landsat 8 Operational Land Imager (OLI) cloud-free images at 30 m spatial resolution were obtained from 2013 to 2017 for the blue (483 nm), green (560 nm) and red (660 nm) channels to retrieve AOD using the Simplified Aerosol Retrieval Algorithm (SARA). The Minimum Reflectance Technique (MRT) was applied on the composite of Landsat surface reflectance images to estimate surface reflectance for AOD retrievals. Validation of the high-resolution AOD retrievals was conducted against two AERONET (AErosol RObotic NETwork) sites located at bright urban surfaces of Beijing. The quality and error of the AOD retrievals were reported using the Expected Error $\mathrm{EE} =\pm(0.05+20\%))$ , root mean square error (RMSE), and relative percent mean error (RPME). Results showed that the Landsat 8 AOD retrievals were well-correlated $\mathrm{R}=0.98$ ) with the AERONET AOD measurements for each channel with 75% of the retrievals were within the EE, RMSE of 0.11 and RPME of 7%. These results suggest that the SARA algorithm is robust and useful to retrieve high-resolution AOD observations and Landsat 8 can be used for aerosols monitoring over bright urban surfaces of Beijing, which is frequently affected by severe dust storms and haze pollution, to identify their effects on public health.