Abstract Retrieval of ice cloud properties from passive remote satellite instruments using multispectral observations of the upwelling radiance at the top of atmosphere is challenging. This requires an accurate separation of the surface spectral reflectance below the cloud and that coming from the ice cloud particle. In this paper, a weighting-function based method is described, which retrieves ice cloud optical thickness (COT) and ice cloud crystal effective radius (CER). The retrieval algorithm belongs to the newly developed eXtensible Bremen Aerosol Retrieval (XBAER) algorithm family, and is called XBAER-ice. The XBAER-ice retrieval uses a surface reflectance parameterization minimizing the impact of surface scattering on the retrieval of optically thin ice cloud properties. A fractal model is used to describe the ice cloud optical properties. The comparison of COT and CER between XBAER-ice and latest Moderate Resolution Imaging Spectroradiometer (MODIS) cloud products shows good agreement. We note that ice cloud property retrievals are particularly sensitive to assumptions about the ice particle shape. The comparison between satellite derived COTs using both the MODIS C6 product and XBAER-ice algorithm agree well with Atmospheric Radiation Measurement (ARM) in-situ measurements (with correlation coefficient above 0.9) whereas CERs show poorer agreement (~0.6). The global spatial distribution patterns and values of COT from ice clouds retrieved by using the MODIS C6 and XBAER-ice algorithms show a high degree of consistency. The XBAER-ice derived ice cloud CER values are systematically smaller than the MODIS C6 product, but both products have similar spatial distribution patterns. The XBAER-ice offers a potential alternative algorithm for the synergistic operational retrieval of ice cloud properties from the observations of Ocean and Land Colour Instrument (OLCI) and Sea and Land Surface Temperature Radiometer (SLSTR) onboard Sentinel-3.

中文翻译：

---

使用改进的 XBAER 算法从反射率的多光谱卫星观测中反演冰云参数

摘要 利用大气顶部上升流辐射的多光谱观测从无源远程卫星仪器中检索冰云特性具有挑战性。这需要准确分离云层下方的表面光谱反射率和来自冰云粒子的表面光谱反射率。在本文中，描述了一种基于加权函数的方法，该方法可以反演冰云光学厚度 (COT) 和冰云晶体有效半径 (CER)。该检索算法属于新开发的可扩展不来梅气溶胶检索（XBAER）算法族，称为XBAER-ice。XBAER-ice 检索使用表面反射参数化最小化表面散射对光学薄冰云特性检索的影响。分形模型用于描述冰云的光学特性。XBAER-ice 与最新的中分辨率成像光谱仪 (MODIS) 云产品之间的 COT 和 CER 比较显示出良好的一致性。我们注意到冰云属性检索对冰粒形状的假设特别敏感。使用 MODIS C6 产品和 XBAER-ice 算法的卫星衍生 COT 之间的比较与大气辐射测量 (ARM) 原位测量非常吻合（相关系数高于 0.9），而 CER 的一致性较差（~0.6）。使用 MODIS C6 和 XBAER-ice 算法反演的冰云 COT 的全球空间分布模式和值显示出高度的一致性。XBAER-ice 导出的冰云 CER 值系统地小于 MODIS C6 产品，但两种产品具有相似的空间分布模式。