All-Day Retrieval of Cloud Physical Properties From Meteosat Second Generation Satellite

The Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard the Meteosat Second Generation (MSG) satellite boasts advantages in continuous cloud monitoring, but its official products suffer from subpar accuracy over challenging surface conditions and insufficient day–night continuity in cloud physical properties. This study develops a near real-time retrieval method for all-day cloud physical properties from SEVIRI thermal infrared (TIR) measurements applying a small attention U-Net (SmaAt-UNet) model. Based on the Moderate Resolution Imaging Spectroradiometer (MODIS) official products, evaluation results show that the overall accuracy (OA) of the cloud phase (CLP) is 87.36%. The root mean square errors (RMSEs) for cloud top height (CTH), cloud effective radius (CER), and cloud optical thickness (COT) are 1.42 km, 4.76 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu $ </tex-math></inline-formula>m, and 10.49, respectively. Compared with the SEVIRI official products, the SEVIRI SmaAt-UNet products display better day–night continuity. To further evaluate the all-day performance of the SmaAt-UNet model over different underlying surfaces, the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) official products were involved. The results demonstrate that the SEVIRI SmaAt-UNet products attain superior all-day retrieval accuracy and better performance than the SEVIRI official products over land. Furthermore, the SEVIRI SmaAt-UNet products are more representative of characterizing the spatial distribution of cloud physical properties thanks to their high spatiotemporal continuity.