Comprehensive upper-air observation network from 1905 to present

The Copernicus Early Upper Air Subdaily Dataset is designed to contain all available upper air data from radiosondes
and PILOT balloons prior to 1979. It is organized in station time series of records on significant levels,
where they can be on height or pressure levels in the source datasets, and where they are on significant pressure
levels in the merged datasets. Main data sources are the NCAR and NCEI upper air datasets as well as data collected at ECMWF during operational assimilation activities. Data collected during the ERA-CLIM and ERA-CLIM2 projects are another important data source.
The source datasets are the basis for a merged dataset, which is subject to strict quality control in order to choose the best available source in case of duplicates. The merged dataset is then homogenized in order to reduce systematic errors, which helps to reduce also the overall observation errors. These are estimated using Desroziers’ method if background and analysis departure information from reanalyses available. The homogenization method is mainly based on analyzing background departure time series from reanalyses, as established in the literature. Users can download the source data, the merged and adjusted merged data, as well as the bias and uncertainty estimates. While data on significant levels are not that convenient to use due to their varying record length, they are uniquely suited for calculating climatologies of e.g. tropopause height, Convective Available Potential Energy or precipitable water.
The merged dataset is supposed to be the cleanest and most complete upper air dataset available for assimilation before 1979 but also for climate research. Data for the period after 1979 are available as well, since they are needed for the used homogenization method anyway, and can be compared with harmonized upper air data as provided by the C3S 311a Lot3 service.
The data content is also available in more integrated form in the companion monthly dataset , which contains gridded fields calculated from the station data, which can be directly compared to other gridded data such as reanalyses or satellite products.
The user should also be aware of an additional historical radiosonde intercomparison dataset ,which provides unique opportunities to diagnose systematic errors between radiosonde types.

哥白尼早期高空次日数据集旨在收录1979年之前所有可用的无线电探空仪和PILOT气象气球高空数据。该数据集以站点时间序列的形式组织，记录在源数据集中显著的高度或气压层，并在合并数据集中集中于显著的气压层。主要数据来源包括美国国家大气研究中心（NCAR）和国家环境信息中心（NCEI）的高空数据集，以及欧洲中期天气预报中心（ECMWF）在业务同化活动期间收集的数据。ERA-CLIM和ERA-CLIM2项目收集的数据也是另一重要数据来源。源数据集是构建合并数据集的基础，合并数据集需经过严格的质量控制，以确保在存在重复数据时选择最佳来源。随后，合并数据集经过同化处理，以减少系统误差，这也有助于降低整体观测误差。这些误差的估算采用Desroziers方法，前提是可获得再分析背景和分析偏差信息。同化方法主要基于分析再分析背景偏差时间序列，如文献所述。用户可以下载源数据、合并数据和调整后的合并数据，以及偏差和不确定性估计。尽管由于记录长度不一，显著层的数据使用不太方便，但它们非常适合计算如平流层高度、对流性可用潜热或降水量的气候学。该合并数据集旨在成为1979年之前最干净、最完整的高空数据集，不仅适用于同化，也适用于气候研究。1979年之后的时期数据也可用，因为它们对于所用的同化方法而言是必需的，并且可以与C3S 311a Lot3服务提供的协调一致的高空数据进行比较。数据内容还以更集成的形式存在于配套的月度数据集中，该数据集包含从站点数据计算出的网格化字段，可以直接与其他网格化数据如再分析或卫星产品进行比较。用户还应注意一个额外的历史无线电探空仪互比数据集，该数据集为诊断不同类型无线电探空仪之间的系统误差提供了独特的机会。