A new cloud detection algorithm based on brightness temperature variation for FY-3C Microwave Humidity Sounder over land

Cloud detection is an important pre-processing procedure in satellite data assimilation. It is necessary to distinguish cloudy and clear-sky data for both clear-sky and all-sky assimilation. However, cloud detection for microwave humidity sounders over land is still challenging due to the variety of surface emissivity and microwave sounders' ability to penetrate clouds. Therefore, a new cloud detection algorithm is proposed for FY-3C microwave humidity sounder Ⅱ (MWHS Ⅱ) data assimilation over land in this study. By using the Community Radiative Transfer Model (CRTM), brightness temperatures of all sounding channels are simulated under different cloud parameters, based on which observations of 7 out of 15 channels are then selected to develop a new land index for cloud detection over land regions. Simulations show that the variety of brightness temperature observations among MWHS-Ⅱ reduces with the increase in cloud height and cloud water content. Compared with AHI cloud products, this new cloud detection method can remove most of the radiances contaminated by clouds, the probability distribution of O-B is more consistent with Gaussian distribution for data on clear sky. This new method has a better capability for identifying supercooled water clouds, opaque ice clouds and overlapping than for cirrus and water clouds with low cloud heights. The detectable rates of supercooled water clouds, opaque ice clouds and overlapping are up to 80%. This algorithm can detect cloudy radiances using MWHS-Ⅱ itself and has a good application prospect.