Evaluation of SGLI/GCOM-C observed cloud properties using surface-based sky radiometer and space-based MODIS and AHI cloud products

The Second-generation Global Imager (SGLI) aboard the Global Change Observation Mission-Climate (GCOM-C, "Shikisai" in Japanese) is an optical sensor observing Earth's surface and atmosphere. The GCOM-C satellite was launched on December 23, 2017. The SGLI observations are expected to improve scientific understanding of climate change mechanism through long-term monitoring of aerosols, clouds, vegetation, and temperature. Among several SGLI products, cloud products have important implications on several research fields including climate change, hydrology, clean energy etc. Thus the qualities of those SGLI observed cloud products need to be understood in detail by comparing them with data of different platforms. We evaluated SGLI observed cloud properties by comparing them with cloud properties observed from surface-based sky radiometer and space-based MODIS and AHI sensors for both water and ice clouds. The surface-observation data of five Japanese SKYNET sites (Chiba, Hedo-misaki, Fukue-jima, Miyako-jima, and Sendai) for the period of January, 2018 to October, 2019 and MODIS and AHI observations over 20°N-40°N, 100°E-140°E on June 24, 2018 are used in this study. Our results show a good agreement between sky radiometer and SGLI observed cloud optical depth (COD) values with relative error less than ~30% for both water and ice clouds. Similarly, SGLI CODs are also found to agree with MODIS and AHI CODs by showing relative errors less than ~14% and ~20%, respectively for both water and ice clouds existing over both land and ocean surfaces. Comparatively, a better agreement between SGLI and each of those reference satellite sensors is found for water clouds and for clouds existing over ocean surface. On the other hand, cloud particle effective radius (CER) values are found to have relatively large differences than COD values for both water and ice clouds for both surface- and space-based platforms with relative error less than ~60%. In general, CER differences are found to be large for ice clouds and for clouds over land.