Clouds are a key modulator of the Earth energy budget at the top of the atmosphere and at the surface. While the cloud top height is operationally retrieved with global coverage, only few methods have been proposed to determine cloud base height (z_base) from satellite measurements. This study presents a new approach to retrieve cloud base heights using the Multi-angle Imaging SpectroRadiometer (MISR) on the Terra satellite. It can be applied if some cloud gaps occur within the chosen distance of typically 10 km. The MISR cloud base height (MIBase) algorithm then determines z_base from the ensemble of all MISR cloud top heights retrieved at a 1.1 km horizontal resolution in this area. MIBase is first calibrated using 1 year of ceilometer data from more than 1500 sites within the continental United States of America. The 15th percentile of the cloud top height distribution within a circular area of 10 km radius provides the best agreement with the ground-based data. The thorough evaluation of the MIBase product z_base with further ceilometer data yields a correlation coefficient of about 0.66, demonstrating the feasibility of this approach to retrieve z_base. The impacts of the cloud scene structure and macrophysical cloud properties are discussed. For a 3-year period, the median z_base is generated globally on a 0.25^∘ × 0.25^∘ grid. Even though overcast cloud scenes and high clouds are excluded from the statistics, the median z_base retrievals yield plausible results, in particular over ocean as well as for seasonal differences. The potential of the full 16 years of MISR data is demonstrated for the southeast Pacific, revealing interannual variability in z_base in accordance with reanalysis data. The global cloud base data for the 3-year period (2007–2009) are available at https://doi.org/10.5880/CRC1211DB.19.

中文翻译：

---

从多角度卫星数据中检索云底高度

云是大气层顶部和地表地球能量收支的关键调节器。虽然可以从全球范围内获取云顶高度，但仅提出了几种方法来根据卫星测量结果确定云底高度（z _base）。这项研究提出了一种使用Terra卫星上的多角度成像光谱辐射仪（MISR）检索云基高度的新方法。如果在选定的通常为10  km的距离内出现一些云隙，则可以应用此方法。然后，MISR云基准高度（MIBase）算法从以1.1检索到的所有MISR云最高高度的集合中确定z _base 该区域的km水平分辨率。首先，使用来自美国大陆上1500多个站点的1年云高仪数据对MIBase进行了校准。在 半径为10 km的圆形区域内，云顶高度分布的第15个百分位数与地面数据提供了最佳一致性。用进一步的云高仪数据对MIBase产品z_基进行全面评估，得到的相关系数约为0.66，说明了这种方法检索z_基的可行性。讨论了云场景结构和宏观物理云属性的影响。对于3年的时间，平均ž_基地全局产生在0.25 ^∘  ×  0.25 ^∘网格。即使阴浊场景和高云从统计排除在外，平均 ž_基地检索过洋以及季节性差异产生似是而非的结果，尤其如此。在东南太平洋地区显示了MISR完整的16年数据的潜力，根据再分析数据，揭示了z_底的年际变化。可以在https://doi.org/10.5880/CRC1211DB.19上获得3年期（2007-2009）的全球云基础数据。