The oceans mitigate climate change by absorbing both atmospheric CO2 corresponding to a significant fraction of anthropogenic CO2 emissions (Friedlingstein et al., 2019; Gruber et al., 2019) and most of the excess heat in the Earth system caused by the enhanced greenhouse effect (Cheng et al., 2020, 2017). The objective of GLODAP (Global Ocean Data Analysis Project; https://www.glodap.info, last access: 29 June 2021​​​​​​​) is to ensure provision of high-quality and bias-corrected water column bottle data from the ocean surface to bottom. These data document the state and the evolving changes in physical and chemical ocean properties, e.g., the inventory of the excess CO2 in the ocean, natural oceanic carbon, ocean acidification, ventilation rates, oxygen levels, and vertical nutrient transports (Tanhua et al., 2021). The core quality controlled and bias-adjusted variables of GLODAP are salinity, dissolved oxygen, inorganic macronutrients (nitrate, silicate, and phosphate), seawater CO2 chemistry variables (dissolved inorganic carbon – TCO2; total alkalinity – TAlk; and pH on the total H+ scale), and the halogenated transient tracers chlorofluorocarbon-11 (CFC-11), CFC-12, CFC-113, and CCl4.

Other chemical tracers that are usually measured on the cruises were included in GLODAP. In many cases a subset of these data is distributed as part of the product; however such data have not been extensively quality controlled or checked for measurement biases in this effort. For some of these variables better sources of data exist, for example the product by Jenkins et al. (2019) for helium isotope and tritium data. GLODAP also includes some derived variables to facilitate interpretation, such as potential density anomalies and apparent oxygen utilization (AOU). A full list of variables included in the product is provided in Table 1.

The oceanographic community largely adheres to principles and practices for ensuring open access to research data, such as the FAIR (Findable, Accessible, Interoperable, Reusable) initiative (Wilkinson et al., 2016), but the plethora of file formats and different levels of documentation, combined with the need to retrieve data on a per cruise basis from different access points, limit the realization of their full scientific potential. In addition, the manual data retrieval is time consuming and prone to data handling errors (Tanhua et al., 2021). For biogeochemical data there is the added complexity of different levels of standardization and calibration and even different units used for the same variable such that the comparability between datasets is often poor. Standard operating procedures have been developed for some variables (Dickson et al., 2007; Hood et al., 2010; Becker et al., 2020), and certified reference materials (CRMs) exist for seawater TCO2 and TAlk measurements (Dickson et al., 2003) and for nutrients in seawater (CRMNS; Aoyama et al., 2012; Ota et al., 2010). Despite this, biases in data still occur. These can arise from poor sampling and preservation practices, calibration procedures, instrument design, and inaccurate calculations. The use of CRMs does not by itself ensure accurate measurements of seawater CO2 chemistry (Bockmon and Dickson, 2015), and the CRMNS have only become available recently and are not universally used. For salinity and oxygen, lack of calibration of the data from conductivity–temperature–depth (CTD) profiler mounted sensors is an additional and widespread problem, particularly for oxygen (Olsen et al., 2016). For halogenated transient tracers, uncertainties in standard gas composition, extracted water volume, and purge efficiency typically provide the largest sources of uncertainty. In addition to bias, occasional outliers occur. In rare cases poor precision – many multiples worse than that expected with current measurement techniques – can render a set of data of limited use. GLODAP deals with these issues by presenting the data in a uniform format, including any metadata either publicly available or submitted by the data originator, and by subjecting the data to primary and secondary quality control assessments, focusing on precision and consistency, respectively. The secondary quality control focuses on deep data, in which natural variability is minimal. Adjustments are applied to the data to minimize cases of bias that could be confidently established relative to the measurement precision for the variables and cruises considered. Key metadata are provided in the header of each data file, and full cruise reports submitted by the data providers are accessible through the GLODAPv2 cruise summary table (https://www.ncei.noaa.gov/access/ocean-carbon-data-system/oceans/GLODAPv2_2021/cruise_table_v2021.html, last access: 7 July 2021).

GLODAPv2.2021 builds on earlier synthesis efforts for biogeochemical data obtained from research cruises, GLODAPv1.1 (Key et al., 2004; Sabine et al., 2005), Carbon dioxide in the Atlantic Ocean (CARINA) (Key et al., 2010), Pacific Ocean Interior Carbon (PACIFICA) (Suzuki et al., 2013), and notably GLODAPv2 (Olsen et al., 2016). GLODAPv1.1 combined data from 115 cruises with biogeochemical measurements from the global ocean. The vast majority of these were the sections covered during the World Ocean Circulation Experiment and the Joint Global Ocean Flux Study (WOCE/JGOFS) in the 1990s, but data from important “historical” cruises were also included, such as from the Geochemical Ocean Sections Study (GEOSECS), Transient Traces in the Ocean (TTO), and South Atlantic Ventilation Experiment (SAVE). GLODAPv2 was released in 2016 with data from 724 scientific cruises, including those from GLODAPv1.1, CARINA, and PACIFICA, as well as data from 168 additional cruises. A particularly important source of data was the cruises executed within the framework of the “repeat hydrography” program (Talley et al., 2016), instigated in the early 2000s as part of the Climate and Ocean – Variability, Predictability and Change (CLIVAR) program and since 2007 organized as the Global Ocean Ship-based Hydrographic Investigations Program (GO-SHIP) (Sloyan et al., 2019). GLODAPv2 is now updated regularly using the “living data process” of Earth System Science Data to document significant additions and changes to the dataset.

There are two types of GLODAP updates: full and intermediate. Full updates involve a reanalysis, notably crossover and inversion, of the entire dataset (both historical and new cruises) and all data points are subject to potential adjustment. This was carried out for GLODAPv2. For intermediate updates, recently available data are added following quality control procedures to ensure their consistency with the cruises included in the latest GLODAP release. Except for obvious outliers and similar types of errors (Sect. 3.3.1), the data included in previous releases are not changed during intermediate updates. Additionally, the GLODAP mapped climatologies (Lauvset et al., 2016) are not updated for these intermediate products. A naming convention has been introduced to distinguish intermediate from full product updates. For the latter the version number will change, while for the former the year of release is appended. The exact version number and release year (if appended) of the product used should always be reported in studies rather than making a generic reference to GLODAP.

Creating and interpreting inversions, as well as other checks of the full dataset needed for full updates, are too demanding in terms of time and resources to be performed every year or 2 years. The aim is to conduct a full analysis (i.e., including an inversion) again after the third GO-SHIP survey has been completed. This completion is currently scheduled for 2023, and we anticipate that GLODAPv3 will become available a few years thereafter. In the interim, the third intermediate update is presented here, which adds data from 43 new cruises to the last update, GLODAPv2.2020 (Olsen et al., 2020).

GLODAPv2.2021 contains data from 989 cruises covering the global ocean from 1972 to 2020, compared to 946 for the period 1972–2019 for GLODAPv2.2020 (Olsen et al., 2020). Information on the 43 cruises added to this version is provided in Table A1 in the Appendix. Cruise sampling locations are shown alongside those of GLODAPv2.2020 in Fig. 1, while the coverage in time is shown in Fig. 2. Not all cruises have data for all of the above-mentioned 12 core variables. For example, cruises with only seawater CO2 chemistry or transient tracer data are still included even without accompanying nutrient data due to their value towards computation of, for example, carbon inventories. In some other cases, cruises without any of these properties measured were included – this was because they did contain data for other carbon-related tracers such as carbon isotopes, with the main intention of ensuring their wider availability. The added cruises are from the years 1982–2020, with most being more recent than 2014. In the Arctic Ocean there are seven cruises from the Canadian Basin carried out on RV Louis S. St-Laurent and one in the Nordic Seas carried out on RV Johan Hjort. In the Pacific Ocean the majority of added cruises are occupations of Line P carried out on RV John P. Tully, as well as a recent occupation of P06 (two legs with different expedition codes, EXPOCODEs) on RV Nathaniel T. Palmer. Note that for some Line P cruises only stations with seawater CO2 chemistry data have been included in the product. Thus, all new Pacific Ocean cruises have seawater CO2 chemistry data. Four out of six cruises added in the Atlantic Ocean (06M220140607 and 06M220160331 on RV Maria S. Merian and 06MT20180213 and 06MT20160828 on RV Meteor) do not have seawater CO2 chemistry data but are included for their transient tracer data. Five new Indian Ocean cruises are added, including the first occupation of GO-SHIP line I07N since 1995. All new cruises from the Indian Ocean include seawater CO2 chemistry data, including pH on three of them, and transient tracers on all (with the exception of a 1982 cruise in the Red Sea on board the RV Marion Dufresne). Finally, three new cruises are added from the Southern Ocean. All of these include seawater CO2 chemistry.

All new cruises were subjected to primary (Sect. 3.1) and secondary (Sect. 3.2) quality control (QC). These procedures are the same as for GLODAPv2.2020, aiming to ensure the consistency of the data from the 43 new cruises with the previous release of this data product (in this case, the GLODAPv2.2020 adjusted data product).

For GLODAPv2.2021 we have also added a basin identifier to the product files, where 1 is the Atlantic Ocean, 4 is the Arctic Mediterranean Sea, 8 is the Pacific Ocean, and 16 is the Indian Ocean. These regions are abbreviated AO, AMS, PO, and IO, respectively, in the adjustment table. Data in the Mediterranean Sea are classified as AO. The basin identifier is now added to the product files to make it easier for users to identify in which ocean basin an individual cruise belongs without having to use one of the four regional files. Note that there is no overlap between the regional files nor our basin identifiers, and cruises in the Southern Ocean are placed in the region where most of the data were collected. In this update we have also included the DOI for each cruise in all product files, with the aim of easing access to the original data and metadata, as well as improving the visibility of data providers.

All material produced during the secondary QC is available via the online GLODAP adjustment table hosted by GEOMAR, Kiel, Germany, at https://glodapv2-2021.geomar.de/ (last access: 29 June 2021), which can also be accessed through https://www.glodap.info. This is similar in form and function to the GLODAPv2 adjustment table (Olsen et al., 2016) and includes a brief written justification for any adjustments applied.

The GLODAPv2.2021 merged and adjusted data product is archived at NOAA NCEI at 10.25921/ttgq-n825 (Lauvset et al., 2021). These data and ancillary information are also available via our web pages https://www.glodap.info and https://www.ncei.noaa.gov/access/ocean-carbon-data-system/oceans/GLODAPv2_2021/ (last access: 7 July 2021). The data are available as comma-separated ascii files (*.csv) and as binary MATLAB files (*.mat) that use the open-source Hierarchical Data Format version 5 (HDF5). The data product is also made available as an Ocean Data View (ODV) file which can be easily explored using the “webODV Explore” online data service (https://explore.webodv.awi.de/, last access: 7 July 2021). Regional subsets are available for the Arctic, Atlantic, Pacific, and Indian oceans. There are no data overlaps between regional subsets, and each cruise exists in only one basin file even if data from that cruise crosses basin boundaries. The station locations in each basin file are shown in Fig. 8. The product file variables are listed in Table 1. A lookup table for matching the EXPOCODE of a cruise with GLODAP cruise number is provided with the data files, and a similar table is provided for matching the GLODAP cruise number with the data DOI. In the MATLAB files this information (EXPOCODE and DOI) is available as a cell array. A “known issues document” accompanies the data files and provides an overview of known errors and omissions in the data product files. It is regularly updated, and users are encouraged to inform us whenever any new issues are identified. It is critical that users consult this document whenever the data products are used.

The original cruise files, with updated flags determined during additional primary GLODAP QC, are available through the GLODAPv2.2021 cruise summary table (CST) hosted by NOAA NCEI: https://www.ncei.noaa.gov/access/ocean-carbon-data-system/oceans/GLODAPv2_2021/cruise_table_v2021.html (last access: 7 July 2021). Each of these files has been assigned a DOI, which is included in the data product files but not listed here. The CST also provides brief information on each cruise and access to metadata, cruise reports, and its adjustment table entry.

While GLODAPv2.2021 is made available without any restrictions, users of the data should adhere to the fair data use principles: for investigations that rely on a particular (set of) cruise(s), recognize the contribution of GLODAP data contributors by at least citing the articles where the data are described and, preferably, contacting principal investigators for exploring opportunities for collaboration and co-authorship. To this end, relevant articles and principal investigator names are provided in the cruise summary table. Contacting principal investigators comes with the additional benefit that the principal investigators often possess expert insight into the data and/or specific region under investigation. This can improve scientific quality and promote data sharing.

This paper should be cited in any scientific publications that result from usage of the product. Citations provide the most efficient means to track use, which is important for attracting funding to enable the preparation of future updates.

GLODAPv2.2021 is an update of GLODAPv2.2020. Data from 43 new cruises have been added to supplement the earlier release and extend temporal coverage by 1 year. GLODAP now includes 47 years, 1972–2020, of global interior ocean biogeochemical data from 989 cruises.

The total number of data records is 1 334 269. Records with measurements for all 12 core variables (salinity, oxygen, nitrate, silicate, phosphate, TCO2, TAlk, pH, CFC-11, CFC-12, CFC-113, and CCl4) are very rare; only 2029 records have measured data for all 12 in the merged product file (interpolated and calculated data excluded). Requiring only two out of the four measured seawater CO2 chemistry variables, in addition to all the other core variables, brings the number of available records up to 9231, and so this is also very rare. A major limiting factor to having all core variables is the simultaneous availability of data for all four transient tracer species: only 26 137 records have measurements of CFC-11, CFC-12, CFC-113, and CCl4 while 422 029 have data for at least one of these (not considering availability of other core variables). A total of 423 544 records have measured data for two out of the three CO2 chemistry core variables. The number of records of measured fCO2 data is 33 844; note again that these data were not subjected to quality control. The number of records with measured data for salinity, oxygen, and nutrients is 832 566, while the number of records with salinity and oxygen data is 1 127 477. All of the above numbers concern measured data, not interpolated or calculated values. A total of 2 % (27 538) of the total data records do not have salinity. There are several reasons for this, the main one being the inability to vertically interpolate due to a separation that is too large (Sect. 3.3.2) between measured samples. Other reasons for missing salinity include salinity not being reported and missing depth or pressure. Note that there are slightly fewer records with fCO2 and all CFC data in GLODAPv2.2021 compared to GLODAPv2.2020. This is due to the removal of data with missing temperatures (Sect. 3.3.1).

Figure 9 illustrates the seasonal distribution of the data. As for previous versions there is a bias around summertime in the data in both hemispheres; most data are collected during April through November in the Northern Hemisphere, while most data are collected during November through April in the Southern Hemisphere. These tendencies are strongest for the poleward regions and reflect the harsh conditions during winter months which make fieldwork difficult. Figure 10 illustrates the distribution of data with depth. The upper 100 m is the best-sampled part of the global ocean, both in terms of number (Fig. 10a) and density (Fig. 10b) of observations. The number of observations steadily declines with depth. In part, this is caused by the reduction in ocean volume towards greater depths. Below 1000 m the density of observations stabilizes and even increases between 5000 and 6000 m; the latter is a zone where the volume of each depth surface decreases sharply (Weatherall et al., 2015). In the deep trenches, i.e., areas deeper than ∼ 6000 m, both number and density of observations are low.

Except for salinity and oxygen, the core data were collected exclusively through chemical analyses of collected water samples. The data of the 12 core variables were subjected to primary quality control to identify questionable or bad data points (outliers) and secondary quality control to identify systematic measurement biases. The data are provided in two ways: as a set of individual exchange-formatted original cruise data files with assigned WOCE flags and as globally and regionally merged data product files with adjustments applied to the data according to the outcome of the consistency analyses. Importantly, no adjustments were applied to data in the individual cruise files, while primary-QC changes were applied.

The consistency analyses were conducted by comparing the data from the 43 new cruises to the previous data product GLODAPv2.2020. Adjustments were only applied when the offsets were believed to reflect biases relative to the earlier data product release related to measurement calibration and/or data handling practices, and not to natural variability or anthropogenic trends. The adjustment table at https://glodapv2-2021.geomar.de/ (last access: 29 June 2021) lists all applied adjustments and provides a brief justification for each. The consistency analyses rely on deep ocean data (> 1500 or 2000 dbar depending on region), but supplementary CANYON-B and CONTENT analyses consider data below 500 dbar. Data consistency for cruises with exclusively shallow sampling was not examined. All new pH data for this version were comprehensively reviewed using crossover analysis, and none required adjustment. Regardless, full reanalysis of all available pH data, particularly in the North Pacific, will be conducted for GLODAPv3.

Secondary QC flags are included for the 12 core variables in the product files. These flags indicate whether (1) or not (0) the data successfully received secondary QC. A secondary QC flag of 0 does not by itself imply that the data are of lower quality than those with a flag of 1. It means these data have not been as thoroughly checked. For δ13C, the QC results by Becker et al. (2016) for the North Atlantic were applied, and a secondary QC flag was therefore added to this variable.

The primary WOCE QC flags in the product files are simplified (e.g., all questionable and bad data were removed). For salinity, oxygen, and the nutrients, any data flagged 0 are interpolated rather than measured. For TCO2, TAlk, pH, and fCO2 any data flags of 0 indicate that the values were calculated from two other measured seawater CO2 variables. Finally, while questionable (WOCE flag = 3) and bad (WOCE flag = 4) data have been excluded from the product files, some may have gone unnoticed through our analyses. Users are encouraged to report on any data that appear suspicious.

Based on the initial minimum adjustment limits and the improvement of the consistency resulting from the adjustments (Table 7), the data subjected to consistency analyses are believed to be consistent to better than 0.005 in salinity, 1 % in oxygen, 2 % in nitrate, 2 % in silicate, 2 % in phosphate, 4 µmolkg-1 in TCO2, 4 µmolkg-1 in TAlk, and 5 % for the halogenated transient tracers. For pH, the consistency among all data is estimated as 0.01–0.02, depending on region. As mentioned above, the included fCO2 data have not been subjected to quality control; therefore no consistency estimate is given for this variable. This should be conducted in future efforts.