北极阿拉斯加站点云观测数据集（1999-2009）

北极阿拉斯加站点云特征集合数据是基于国际上5种著名云反演产品的融合集成数据，数据覆盖时间从1999年到2009年，时间分辨率为逐小时，垂直层数为512层，垂直分辨率45米，覆盖站点为北极阿拉斯加站点，经纬度坐标为（71°19′22.8″N, 156°36′32.4″ W）。所包含云遥感反演数据产品包括：美国能源部大气辐射观测计划采用参数化方法遥感反演获得的所有相态云特征官方产品，美国NOAA Matt Shupe和DaveTurner合作遥感反演（最优化方法+参数化方法）获得的冰云和混合云特征产品，美国University of Wyoming大学Zhien Wang获得的混合云（最优化方法）特征产品，美国University of Wyoming大学Min Deng获得的冰云（参数化方法）特征产品，以及美国State University of New York at Albany大学Qilong Min遥感反演（最优化方法）获得的云光学厚度产品。遥感产品变量包含云水有效半径、云水含量、云冰有效半径、云冰含量、云光学厚度、云水柱含量；相应的观测反演误差范围约为10-30%，30-60%，10-30%，30-60%，10-30%和10-20%。数据格式为nc格式，每月一个nc文件。

The Cloud Feature Ensemble Dataset at the Arctic Alaska Site is a fused and integrated product based on 5 well-known international cloud retrieval products. The dataset covers the period from 1999 to 2009, with an hourly temporal resolution, 512 vertical levels, and a vertical resolution of 45 meters. The targeted observation site is the Arctic Alaska station, with geographic coordinates of (71°19′22.8″N, 156°36′32.4″W). The included cloud remote sensing retrieval products are as follows: 1) the official all-phase cloud feature product retrieved via parametric method by the U.S. Department of Energy Atmospheric Radiation Measurement (ARM) Program; 2) the ice cloud and mixed-phase cloud feature product retrieved via combined optimal estimation and parametric methods by Matt Shupe and Dave Turner from the U.S. National Oceanic and Atmospheric Administration (NOAA); 3) the mixed-phase cloud feature product retrieved via optimal estimation method by Zhien Wang from the University of Wyoming; 4) the ice cloud feature product retrieved via parametric method by Min Deng from the University of Wyoming; and 5) the cloud optical thickness product retrieved via optimal estimation method by Qilong Min from the State University of New York at Albany. The variables of these remote sensing products include cloud water effective radius, cloud water content, cloud ice effective radius, cloud ice content, cloud optical thickness, and cloud liquid water path. The corresponding retrieval error ranges are approximately 10–30%, 30–60%, 10–30%, 30–60%, 10–30%, and 10–20%, respectively. The data is stored in NetCDF (nc) format, with one nc file per month.