Cloud Removal Algorithm by Combining MODIS and VIIRS Snow Products

The moderate resolution imaging spectroradiometer (MODIS) and the visible infrared imaging radiometer suite (VIIRS) sensors share similar spectral band configurations and are widely utilized for snow mapping due to their high spatial and moderate temporal resolution. However, differences in cloud identification algorithms lead to significant discrepancies in cloud and snow distribution between MODIS and VIIRS snow products. This article compares three snow products—MOD10A1, MYD10A1, and VNP10A1—and proposes a cloud removal method by integrating these products. The main findings are as follows: First, across China’s three major snow-covered regions, the cloud coverage in MODIS snow products consistently exceeds that of VIIRS snow products. Second, in the high-latitude forested areas of Northeast China, numerous snow-covered pixels are accurately identified in the VIIRS snow product but are misclassified as clouds in the MODIS snow products. Third, the proposed cloud removal algorithm, which integrates MODIS and VIIRS data, effectively leverages the complementary information from the three snow products, reducing the average cloud coverage from more than 40% to less than 10%. This represents a significant improvement compared to the standalone MODIS and VIIRS products.