Retrieving Aerosol Optical Depth over Land from Landsat-8 Satellite Images with the Aid of Cloud Shadows

Clouds and their shadows can be clearly identified from high-spatial-resolution satellite images, such as those provided by Landsat-8/9 with a spatial resolution of approximately 30 m and Sentinel-2 with a spatial resolution of around 20 m. Consequently, the difference between satellite measurements over cloud-shadowed and nearby illuminated pixels can be used to derive the aerosol optical depth (AOD), even in the absence of detailed surface optical properties. Based on this assumption, an algorithm for AOD retrieval over land is developed and tested using Landsat-8/9 images containing scattered clouds over Xuzhou, China, and Dalanzadgad, Mongolia. The retrieved AODs are compared against MODIS and ground-based sun photometer measurements. The findings reveal that, in cloudy regions, over 90% of the discrepancies between the AODs retrieved using the cloud-shadow method and ground-based measurements fall within 0.05 +/- 0.20 AOD. This cloud-shadow algorithm represents a valuable complement to existing satellite aerosol retrieval methods, particularly in sparsely cloud-covered areas.