3S-GEOPROF-COMB: A Global Gridded Dataset for Cloud Vertical Structure from combined CloudSat and CALIPSO observations

Global cloud dataset from combined spaceborne radar and lidar.

This repository contains the 3S-GEOPROF-COMB product, a globally-gridded dataset for cloud vertical structure retrieved from hybrid active remote sensing (CloudSat radar and CALIPSO lidar) reported at 240 m vertical resolution. Science variables include vertical cloud fraction and vertically-integrated cloud cover for various geometrical criteria (i.e. high, middle, low, and thick clouds, along with with unique high, middle, and low cloud cover variants).

A Python notebook showing how to work with the dataset is available on GitHub, as is the source code used to produce the data product.

Our product is calculated from the latest release (R05) of per-orbit (level 2) combined cloud mask profiles in 2B-GEOPROF-LIDAR with additional data from 2B-GEOPROF. Validation and a complete description of the data product is given in the paper "A Global Gridded Dataset for Cloud Vertical Structure from Combined CloudSat and CALIPSO Observations" (Earth System Science Data).

Please cite "Bertrand, L., Kay, J. E., Haynes, J., and de Boer, G.: A global gridded dataset for cloud vertical structure from combined CloudSat and CALIPSO observations, Earth Syst. Sci. Data, 16, 1301–1316, https://doi.org/10.5194/essd-16-1301-2024, 2024."

The files contained in each folder are given via the following format:

          instruments_frequency_resolution.zip

instruments:

radarlidar: the standard product, computed from merged geometrical profiles of hydrometeor occurrence

radaronly: computed solely from CloudSat radar profiles, otherwise processing is identical. For when users need to determine which instrument is responsible for observations of interest.

lidaronly: computed solely from CALIPSO lidar profiles, otherwise processing is identical. For when users need to determine which instrument is responsible for observations of interest.

frequency:

monthly: data files report fields aggregated over a 1-month period

seasonal: data files report fields aggregated over a 3-month period (DJF, MAM, JJA, SON)

resolution:

2.5x2.5: each grid box spans 2.5 degrees latitude and 2.5 degrees longitude

5x5: each grid box spans 5 degrees latitude and 5 degrees longitude

10x10: each grid box spans 10 degrees latitude and 10 degrees longitude

Each folder contains a netCDF data file and a cloud cover quicklook plot image file for each time period over the 2006-2019 data record. Individual files are named according to the following format:

          timeperiod_instruments_datastream_version.nc (or .png)

timeperiod: the time step at the given frequency, either e.g. 2006-08 (August 2006) or 2012-DJF (December 2012 to February 2013).

instruments: the instruments used in the data product as a whole, always CSCAL (CloudSat and CALIPSO).

datastream: either 3S-GEOPROF-COMB (COMBined radar and lidar), 3S-GEOPROF-COMB-RO (the auxiliary Radar Only variant of the product), or 3S-GEOPROF-COMB-LO (the auxiliary Lidar Only variant of the product)

version: current release is v8.4

The product handles the 2011 CloudSat battery anomaly, after which the satellite only collects data in the sunlit portion of its orbit, by allowing users to subsample the pre-anomaly period to mimic the post-anomaly collection patterns. This allows users to estimate the effect of the reduced sampling on their analyses or apply a consistent sampling mode to the entire dataset. This option is provided to users via the "doop" dimension. Dimension coordinate value "All cases" reports variables computed using all observations, while "DO-OP observable" reports variables using only input data that either were or would have been collected in DO-OP mode (i.e. the pre-DO-OP period is subsampled to DO-OP collection patterns).

基于星载雷达与激光雷达融合的全球云数据集

本仓库收录3S-GEOPROF-COMB产品，这是一套通过混合主动遥感技术（CloudSat雷达（CloudSat radar）与CALIPSO激光雷达（CALIPSO lidar））反演得到的全球网格化云垂直结构数据集，垂直分辨率为240米。该数据集的科学变量包含多种几何判据下的垂直云量与垂直积分云覆盖率，涵盖高云、中云、低云与厚云，同时包含专属的高云、中云、低云覆盖率变体。

GitHub平台上提供了用于处理该数据集的Python交互式笔记（Python notebook），以及生成该数据产品的源代码。

本产品基于2B-GEOPROF-LIDAR的逐轨（二级）组合云掩码廓线的最新版本（R05），并结合2B-GEOPROF的额外数据计算得到。该数据产品的验证方法与完整说明可参阅论文《A Global Gridded Dataset for Cloud Vertical Structure from Combined CloudSat and CALIPSO Observations》（发表于《地球系统科学数据（Earth System Science Data）》）。

请引用如下文献：Bertrand L、Kay J E、Haynes J、de Boer G. 基于CloudSat与CALIPSO联合观测的全球云垂直结构网格化数据集[J]. 地球系统科学数据, 2024, 16: 1301-1316. https://doi.org/10.5194/essd-16-1301-2024.

每个文件夹内的文件遵循以下命名格式：`instruments_frequency_resolution.zip`

### 仪器类型（instruments）
- `radarlidar`：标准产品，基于水凝物出现的合并几何廓线计算得到；
- `radaronly`：仅基于CloudSat雷达廓线计算得到，其余处理流程完全一致，适用于用户需要明确目标观测对应仪器来源的场景；
- `lidaronly`：仅基于CALIPSO激光雷达廓线计算得到，其余处理流程完全一致，适用于用户需要明确目标观测对应仪器来源的场景。

### 时间频率（frequency）
- `monthly`：数据文件包含以1个月为周期聚合的统计字段；
- `seasonal`：数据文件包含以3个月为周期聚合的统计字段，分别对应DJF（12月-次年2月）、MAM（3月-5月）、JJA（6月-8月）、SON（9月-11月）。

### 空间分辨率（resolution）
- `2.5x2.5`：每个网格单元覆盖2.5°纬度 × 2.5°经度；
- `5x5`：每个网格单元覆盖5°纬度 × 5°经度；
- `10x10`：每个网格单元覆盖10°纬度 × 10°经度。

每个文件夹包含2006-2019年数据记录中各时间段对应的网络通用数据格式（netCDF）数据文件，以及云覆盖率快速预览图像文件。单个文件的命名格式如下：`timeperiod_instruments_datastream_version.nc`（或`.png`）

各字段含义如下：
- `timeperiod`：对应指定频率的时间步长，格式示例为`2006-08`（2006年8月）或`2012-DJF`（2012年12月至2013年2月）；
- `instruments`：数据产品整体采用的观测仪器，固定为`CSCAL`（CloudSat与CALIPSO）；
- `datastream`：数据流类型，可选值包括`3S-GEOPROF-COMB`（雷达与激光雷达联合产品）、`3S-GEOPROF-COMB-RO`（辅助雷达仅变体产品）以及`3S-GEOPROF-COMB-LO`（辅助激光雷达仅变体产品）；
- `version`：当前最新版本为`v8.4`。

针对2011年CloudSat卫星电池异常事件（此后卫星仅在轨道日照区段收集数据），本产品支持用户通过预异常时段的子采样来模拟后异常时段的采集模式，以此评估采样缩减对分析结果的影响，或对全数据集采用统一的采样模式。该功能通过`doop`维度实现：维度坐标值`"All cases"`代表使用全部观测数据计算得到的变量，`"DO-OP observable"`代表仅使用DO-OP模式下采集（或本可采集）的输入数据计算得到的变量（即预DO-OP时段被子采样至DO-OP采集模式）。