| Summary: | A collection of oxygen data from two different sources. The first source is the bottle data from the Ocean Station Data (OSD) in the World Ocean Database 2018 (WOD18) collection (Boyer et al., 2018). The second source is the sensor data from the BGC Argo floats collection (Argo (2023). The oxygen collection is created to be consistent with the GLODAPV2.2019 database, aiming to reduce the bias caused by systematic errors in the collected oxygen data. This is achieved through a procedure called secondary quality control, which involves correcting the oxygen values towards the GLODAPV2 oxygen data using a cross-over analysis with the MATLAB toolbox prepared by Lauvset and Tanhua (2015) and with the GLODAPV2 data product as reference. Further details on the secondary quality control methodology are provided in Olsen et al. (2019). The oxygen values in the collection are already the corrected values. However, we have also included the adjustment values applied to each individual oxygen data point. These adjustments are multiplicative, meaning that to obtain the original oxygen values, the corrected oxygen value in the data collection must be divided by the correction factor. Additionally, the collection includes temperature and salinity data, though these variables did not undergo the secondary quality control process. Also included in the data collection are the flags from the first quality control, which are directly sourced from the two data repositories. It's worth noting that the flag values from the bottle data in WOD18 differ from the flag values associated with the sensor data from the Argo floats. Users should refer to the respective datasets if they intend to flag bad data. As a general guideline for the WOD18 data, flags other than 0, 2, and 6 should be considered as indicative of bad data. Similarly, for the Argo floats data, values other than 1, 2 (probably good), 5, and 8 should be flagged as potentially bad.
|
|---|