Quality control procedures and methods of the CARINA database
Quality control procedures and methods of the CARINA database
复制标题
CARINA数据库的质量控制程序和方法
DOI：
10.5194/essd-2-35-2010
复制
发表时间：
2009
影响因子：
11.4
通讯作者：
C. Schirnick
中科院分区：
文献类型：
--
作者：
T. Tanhua;S. Heuven;R. Key;A. Velo;A. Olsen;C. Schirnick
抽象的。从以前在北极、大西洋和南大洋进行的不公开的航行中获得的关于碳和与碳有关的水文和水化学参数的数据已被检索出来，并合并到一个新的数据库：CARINA（CARbon IN the Atlantic）。这些数据经过严格的质量控制（QC）程序，以确保最高的质量和一致性。所有CARINA数据都经过初步质量控制;这是一个研究数据以确定离群值和明显错误的过程。此外，对CARINA数据库中的几个测量参数进行了二次QC。二级QC是对数据进行客观研究以量化报告值中的系统差异的过程。这一过程涉及交叉分析，第二步是利用交叉分析得出的偏移量，利用最小二乘模型计算对各次航行测得的参数的修正。在数据产品中纠正了数据中发现的重大偏差，即三个合并的数据文件，其中包含三个区域（即北极地中海、大西洋和南大洋）的测量、计算和插值数据。在这里，我们报告的质量控制的技术细节和工具，已开发和使用的项目，包括交叉分析和最小二乘模型的程序。此外，还开发了一个用于上传结果、绘图、评论等的互动网站，该网站对项目的成功至关重要，本文也对此进行了描述。
Abstract. Data on the carbon and carbon relevant hydrographic and hydrochemical parameters from previously not publicly available cruises in the Arctic, Atlantic and Southern Ocean have been retrieved and merged to a new data base: CARINA (CARbon IN the Atlantic). These data have gone through rigorous quality control (QC) procedures to assure the highest possible quality and consistency. All CARINA data were subject to primary QC; a process in which data are studied in order to identify outliers and obvious errors. Additionally, secondary QC was performed for several of the measured parameters in the CARINA data base. Secondary QC is a process in which the data are objectively studied in order to quantify systematic differences in the reported values. This process involved crossover analysis, and as a second step the offsets derived from the crossover analysis were used to calculate corrections of the parameters measured on individual cruises using least square models. Significant biases found in the data have been corrected in the data products, i.e. three merged data files containing measured, calculated and interpolated data for each of the three regions (i.e. Arctic Mediterranean Seas, Atlantic, and Southern Ocean). Here we report on the technical details of the quality control and on tools that have been developed and used during the project, including procedures for crossover analysis and least square models. Furthermore, an interactive website for uploading of results, plots, comments etc. was developed and was of critical importance for the success of the project, this is also described here.