资源环境科技发展态势分析平台

Despite the vertically inhomogeneous (VIH) structure of ice clouds, current passive remote sensing methods assume plane-parallel homogeneous (PPH) layers, which can lead to retrieval errors. An adequate VIH cloud model is required to improve retrieval performance. In this study, CloudSat and CALIPSO satellite measurements in a 1-year period were analyzed to model cloud vertical inhomogeneity, and its impacts on cloud retrieval were assessed using thermal infrared (TIR) measurements. The satellite measurements revealed that the peak ice water content (IWC) located around the cloud vertical midpoint moved toward the cloud base as the ice water path (IWP) increased in clouds with small IWP values; thicker clouds exhibited a gradual shift in IWC peak location toward the cloud top as IWP increased. The vertical profiles of both the cloud-particle effective radius (CER) and a proxy of cloud-particle number concentration showed close associations with the vertical IWC profile. An empirical model linking cloud geometrical thickness to columnar optical properties (IWP and column-mean CER) as a function of cloud-top temperature was also proposed. Compared with a model assuming PPH clouds, the VIH cloud model improved retrieval performance by reducing the retrieval error of the TIR-based passive remote sensing algorithm. Further, by increasing the retrieved values of cloud-top height noticeably for high-level clouds and column-mean CER considerably, with minimal effects on cloud optical thickness retrieval, the VIH cloud model yielded results that were in better agreement with radar/lidar ice cloud products than those obtained under the assumption of PPH cloud layers.