Multilayer cloud detection and retrieval of cloud physical and optical properties from thermal infrared measurements

Abstract:

Recent studies using active remote sensing have revealed significant occurrence of multi-layer cloud. Detection of multi-layer cloud is important in passive remote sensing for quality assessment of cloud property retrieval and identification of uncertain retrievals. An algorithm using several thermal infrared (TIR) bands at 6–13.5 micron wavelengths to detect multilayer cloud and retrieve cloud physical and optical properties including cloud thermodynamic phase is developed. This significantly extends applicability of passive remote sensing and improves accuracy of cloud property retrieval. The method uses the split window bands as well as the carbon dioxide and water vapor absorption bands. The forward model uses the two-stream approximation to solve radiative transfer with gaseous absorption treated by the correlated-k distribution method. Brightness temperature errors are evaluated by model-to-model and model-to-measurement comparisons. Top pressure of lower cloud in multi-layer cloud column can be retrieved if the upper cloud optical thickness is less than 6. The optimal estimation method is used to simultaneously infer several cloud properties including water path, effective particle radius and cloud-top pressure. The method is applied to the Moderate Resolution Imaging Spectroradiometer (MODIS) using 10 TIR bands and compared to MODIS operational product and active remote sensing measurements, showing promising results. The TIR method well detects optically thin clouds and retrieve their properties with relatively high accuracy. Particularly, cloud-top of optically thin cloud is estimated well. Multi-layer cloud detection works usually, while the TIR measurements miss very thin cloud that appears near the tropopause. The algorithm will be applied to frequent observation data from a new Japanese geostationary satellite, Himawari-8.