the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Jul 2026
Δ14C and δ13C of DIC dataset from R/V Mirai and NABOS in the Arctic Ocean, with a synthesis of published records from 1979–2021
Abstract. We present a new full-water-column dataset of dissolved inorganic carbon (DIC) Δ¹⁴C, δ¹³C, DIC concentration, and CTD/chemical properties for the Arctic Ocean, obtained on R/V Mirai cruise transects (1999, 2000, 2002, 2006, 2008, 2009) and the 2008 NABOS expedition. In total, 255 new DIC Δ¹⁴C measurements (each with paired δ¹³C and DIC concentration) are reported here, of which 42 are from NABOS 2008. The archived dataset released with this study consists primarily of these unanalyzed data; raw numerical values from earlier published papers are not redistributed. To evaluate the long-term change of Arctic DIC Δ¹⁴C over 1979–2021 are used in the text for citation-based comparison and synthesis analysis. Older values transcribed or extracted from printed pages are treated as contextual information for surveying the time series and vertical structure, not as redistributed individual values. The primary aims of this paper are (i) to release the new Mirai/NABOS data in a long-term-archivable form, (ii) to organise the location, period, region, depth range and water-mass information of the published data, and (iii) by combining the new data with published records, to demonstrate the scientific utility of the full-water-column Arctic DIC Δ¹⁴C structure, including the surface mixed layer, the PWW/halocline, the AW layer and the deep water. The detailed interpretation of the temporal change of the AW layer and of Atlantification is deferred to a companion JGR-Oceans paper (Uchida et al., submitted).
- Preprint
(14871 KB) - Metadata XML
-
Supplement
(43 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- CC1: 'Comment on essd-2026-493', Anne-Marie Wefing, 06 Aug 2026
-
RC1: 'Comment on essd-2026-493', Anonymous Referee #1, 24 Aug 2026
General comments
This manuscript presents a valuable set of previously unpublished Arctic DIC radiocarbon and stable carbon isotope observations from R/V Mirai and NABOS cruises and attempts to place these measurements in a longer-term Arctic context. The underlying observations are potentially useful, particularly as full-depth DIC Δ14C observations in the Arctic remain relatively sparse.
However, in its current form, I do not think the manuscript yet meets the standard expected for an ESSD data description paper. The main concern is not the value of the observations, but the definition and completeness of the data product. The manuscript combines a relatively small newly released dataset with a larger synthesis of published observations, while much of the scientific interpretation, particularly for the Atlantic Water (AW) layer and Atlantification, is explicitly assigned to a companion JGR-Oceans manuscript. This creates substantial overlap and makes the specific contribution of the ESSD paper difficult to identify. The manuscript itself states that the detailed AW-layer interpretation is deferred to the companion paper, while nevertheless presenting the four-phase AW evolution, its slopes, and the AW plateau in the present manuscript. This substantially erodes the novelty of the companion JGR-Oceans paper, and raises reasonable concerns about salami slicing of a single research project into two publications. The manuscript relies on 1025 observations compiled from published records for its core synthesis analyses, yet only provides citations and DOIs for these historical data without delivering a harmonized, quality-controlled compilation integrated with the 255 newly released measurements. In addition, the present data release appears to be centered on the 255 newly reported isotope measurements, whereas associated CTD, hydrographic, dissolved oxygen, carbonate-system, and other cruise observations are mainly referenced through external cruise records rather than being incorporated into a single analysis-ready product. This represents a significant missed opportunity for an ESSD contribution.
Major comments
- The manuscript needs a much clearer ESSD-specific purpose.The paper should focus on the creation, quality control, documentation, and long-term usability of an Arctic DIC isotope data product rather than on a partial scientific interpretation of the Arctic Δ14C record. At present, the manuscript contains substantial scientific interpretation of SML, PWW, AW, deep-water age, and Atlantification, while the actual data product remains relatively limited. I recommend substantially reducing the process-oriented interpretation and restructuring the manuscript around the data product itself.
- The relationship between this ESSD manuscript and the companion JGR-Oceans paper must be clarified.The current manuscript presents the four-phase AW evolution, including the increase, plateau, rapid decline, and slow decline, although these results are explicitly described as part of the companion JGR-Oceans paper. The same information is repeated in the Results and Conclusions. This raises a fundamental question: what is the distinct scientific contribution of the JGR-Oceans paper if the major AW temporal features and their quantitative characteristics are already presented here? The two papers should be clearly separated. The ESSD paper should establish the dataset, its provenance, processing, uncertainty, validation, and demonstrated utility.
- The manuscript currently reports 255 new Δ14Cmeasurements with paired δ13C and other information, but associated cruise observations such as CTD, bottle-water chemistry, dissolved oxygen, ｐCO2, nutrients, and other hydrographic and geochemical measurements are only referenced through external repositories. For an ESSD publication, I strongly recommend that the authors go one step further and construct an integrated, analysis-ready product aligned with the standards of widely used community reference datasets such as GLODAP.
- The manuscript uses 1025 observations by combining the new data with published records, but explicitly states that the statistics presented from Table 2 onward are based on a synthesis that combines new and published data, rather than on the archived dataset itself.If these 1025 observations are central to the scientific demonstration of the product, I recommend making the full compilation available in ananalysis-ready form, with complete provenance for each record. The target should be a single harmonized Arctic DIC isotope database containing the full set of publicly reusable observations, rather than a new-data file plus a separate literature inventory. Bringing them together would substantially increase the value and reusability of the dataset. ESSD specifically recommends complete source attribution, processing documentation, validation, uncertainty accounting, and accessible data products and codes.
- The synthesis combines Δ14Crecords from severalsources across 1979–2021, generated by different laboratories, AMS facilities, extraction methods and normalization protocols. The only stated harmonization step is conversion to Stuiver & Polach (1977) notation, which addresses reporting convention but not analytical systematic biases. Direct trend calculation from uncalibrated multi-source data undermines the reliability of reported rates such as -2.89‰ yr-1. A dedicated crossover analysis using overlapping stations and depth horizons is necessary to quantify and mitigate inter-measurement biases. For each data source, the authors should document full provenance and processing details: laboratory and analytical method, reference standards and normalization protocols, originally reported analytical uncertainty, all decay and normalization corrections applied, unit and notation conversion procedures, treatment of missing or estimated values, quality-control criteria, whether individual values were digitized from tables, supplementary files or figures, and whether source-specific uncertainties were retained or reassigned.
- The title gives Δ14Cand δ13Cequal prominence, but the scientific analysis is overwhelmingly based on Δ14C. δ13C is mainly presented as a paired measurement and receives limited analysis. The authors are advised to either substantially expand the description and scientific utility demonstration of the δ13C component throughout the manuscript, or revise the title and overall framing to clearly position δ13C as an auxiliary variable accompanying the primary Δ14C radiocarbon dataset.
- The current abstract emphasizes the scientific synthesis and ends by pointing readers toward the companion JGR-Oceans paper. This weakens the identity of the ESSD paper. Besides, the abstract fails to clearly distinguish between the 255 newly released measurements and the 1025-point synthesis dataset compiled from published literature, which may lead readers to incorrectly assume this paper releases a full 1979–2021 dataset. The abstract should be rewritten around the data product.
Minor Comments
- Several quantitative interpretations should be presented more cautiously. For example, the trend analyses are based on highly uneven temporal and spatial sampling, and the manuscript uses both raw-sample regressions and cast-mean regressions. The statistical justification for these choices is currently insufficient. Confidence intervals, regression uncertainty, and sensitivity to sampling strategy should be reported.
- Some interpretations are stronger than the presented data support. For example, near-zero surface Δ14Cvalues are described as indicating “recent freshening”(Table 3). Δ14C alone does not directly demonstrate freshening, such an interpretation requires independent salinity or freshwater-fraction evidence. Similarly, the interpretation of Δ14C and δ13C changes in the Chukchi shelf as direct evidence of photosynthetic uptake and benthic respiration is plausible but should be stated as being consistent with these processes.
- The manuscript contains redundant phrasing, duplicated expressions and occasional awkward long sentences. For example, the manuscript repeatedly explains that published data are not redistributed. This information should be consolidated into one clear data-policy section rather than repeated throughout the Results. A full round of language editing and conciseness improvement is needed.
- There is an arithmetic inconsistency in the Conclusion: the statement in the Line 500-501 “SML Δ14Cdeclined by about 90‰ over 42 years (1987 peak +89‰ → 2021 −7‰)” is incorrect. The period 1987-2021 is 34 years, not 42 years (which would correspond to 1979–2021).
- The authors appropriately disclose their use of Anthropic Claude for language editing, formatting, translation support, and figure preparation. However, multiple factual errors and textual irregularities throughout the manuscript point to insufficient human validation of AI-processed content. Specifically, the citation of Povinec et al. (2011) as evidence for subarctic North Pacific Δ14Cdecline is misapplied, as that work investigates tracer distributions in the southern Indian Ocean, not the subarctic North Pacific. As noted by Anne-Marie Wefing, the reference entry for Payne et al. (2024) lists incorrect co-authors despite bearing an accurate DOI. Widespread redundant phrasing and repetitive statements also indicate incomplete manual revision of AI-edited prose. The the authors should explicitly verify all scientific content, numerical values, and citations against the original sources.
Citation: https://doi.org/10.5194/essd-2026-493-RC1 -
RC2: 'Comment on essd-2026-493', Anonymous Referee #2, 30 Aug 2026
General comments
This manuscript presents 255 full-water-column DIC Δ¹⁴C observations collected during six R/V Mirai cruises and the NABOS 2008 expedition. These costly and historically fixed measurements could provide a useful resource for future Arctic water-mass studies, model evaluation, and tracer syntheses. The archived workbook is generally well organised. It contains explicit units and source identifiers, and the variable dictionary covers all 24 data columns. I also found no exact duplicate rows, invalid geographic coordinates, negative depths, or negative DIC concentrations.
The data DOI resolves to the NIPR Arctic and Antarctic Data archive System record, version 1.00, with a CC BY licence and English metadata. The data citation is included in both the Data availability section and the reference list. The supplied ZIP contains a flat data workbook and a README, and the files can be inspected without specialised scientific software. Providing CSV companions would further improve machine reuse, but this is not a condition on its own.
Several issues nevertheless limit the present auditability and reuse of the dataset and the larger synthesis. These include a potentially erroneous δ¹³C uncertainty field, inconsistencies between the dataset scope described in the manuscript and the contents of the released files, the absence of the published-data inventory described as documenting the published-data inputs to the synthesis, insufficient uncertainty and quality assessment, conflicting sample counts, and the amount of scientific interpretation relative to the scope of an ESSD data-description article. I therefore recommend major revisions. These concerns should be resolvable through data verification, clearer documentation, narrower claims, and reproducible provenance. I am not requesting new field observations or an expanded scientific analysis.
Major comments
1. Verification of the δ¹³C uncertainty field. The δ¹³C uncertainty field should be verified before publication (p. 9, lines 142–150, and New_Mirai_NABOS_DIC14C_dataset.xlsx, sheet DIC_isotope_data, Excel rows 82–93 and 164–180). In rows 82–93 from MR00-K06 station A00-023_3, Delta13C_error_permil is identical to Delta13C_permil for 12 consecutive samples, which suggests a possible transcription or column-mapping error. In addition, the uncertainties reported for station M0863 in rows 164–180 range from 0.311 to 0.648‰, substantially exceeding the representative precision of ±0.05‰ stated in the manuscript. Please verify both ranges against the source or laboratory records. If corrections are needed, the workbook, README, repository metadata, manuscript, and published-data inventory should be synchronised under a new dataset version, and only analyses that depend on the corrected values need to be rerun. If the values are correct, a short explanation of what the field represents and why these samples have larger uncertainties would resolve the concern.
2. Consistency between the stated dataset scope and the released files. The scope of the released dataset should be described consistently throughout the manuscript (p. 1, lines 17–21, p. 4, lines 82–84, p. 10, lines 158–160, p. 11, lines 177–187, and the Table 7 note on p. 27, lines 462–468). Several passages state that CTD and chemical variables are included or attached, whereas sheet Cruise_sources in New_Mirai_NABOS_DIC14C_dataset.xlsx and the accompanying README direct users to external cruise records rather than redistributing those variables. In addition, in sheet DIC_isotope_data of New_Mirai_NABOS_DIC14C_dataset.xlsx, column DIC_umol_kg is blank in 88 of the 255 data records (Excel rows 2–256), so not every Δ¹⁴C observation has a paired DIC concentration. Please revise the Abstract, Sect. 2.1, Sect. 3, the Table 7 note, the README, and the repository metadata to distinguish variables redistributed in the workbook, variables available only for a subset of observations, and variables available only from external records. For the external CTD and hydrographic data, please identify the source record and the station, cast, sample, date, or depth fields needed to join the records. The external data do not need to be duplicated if the scope and joining procedure are made clear.
3. Auditable provenance for the published-data synthesis. The provenance of the larger synthesis should be made independently auditable (p. 10, lines 161–164, and p. 28, lines 469–483). Figures 1–4 and Tables 2–6 use a new-plus-published compilation, while sheet DIC_isotope_data in New_Mirai_NABOS_DIC14C_dataset.xlsx contains only the 255 new observations. The manuscript states that Published_data_inventory.xlsx records the published sources and is archived with the dataset, but this file is absent from the supplied supplement ZIP. Please provide a direct, version-specific repository link if the inventory is already deposited, or deposit it with the revised materials. The inventory or accompanying workflow should document each source, the inclusion and exclusion criteria, the extraction method, and the subset used in each figure or table. Values that cannot lawfully be redistributed need not be included, but their provenance and selection should still be documented. Where licences permit, an analysis-ready synthesis table and the code used for the figures, tables, and regressions would further improve reproducibility.
4. Dedicated uncertainty and quality assessment. The manuscript should include a dedicated uncertainty and quality-assessment section (p. 9, lines 142–157, p. 10, lines 158–164, and New_Mirai_NABOS_DIC14C_dataset.xlsx, sheet DIC_isotope_data, Excel rows 2–256). The representative analytical precision currently reported does not explain how data quality varies among samples, cruises, analytical batches, reanalysed observations, and legacy records. To the extent supported by the surviving records, please describe the measurement standards and blank or background treatment, define the derivation and statistical meaning of each uncertainty field, distinguish precision from accuracy, explain the treatment of repeats and reanalysed samples, and confirm whether the reported δ¹³C values are independent IRMS measurements across all cruises and samples, identifying any exceptions. No new field campaign or impossible retrospective validation is requested. If independent reference data are unavailable, a quantitative plausibility assessment based on profiles, repeat measurements, source comparisons, and explicit legacy-data limitations would be sufficient.
5. Reconciliation of numerical populations. The manuscript should reconcile the conflicting population sizes and define what each reported N represents. Figure 1 gives N = 59, 56, and 45 in the panel headings on pp. 6–7, whereas its caption on pp. 7–8, lines 100–115, gives N = 51, 49, and 38. Sect. 4 states N = 154 on p. 16, lines 276–283, while Table 3 on p. 18 sums to 138. Sect. 7 states N = 119 on p. 22, lines 381–387, while Table 5 sums to 125. The synthesis population is reported as N = 1025 on p. 11, lines 181–182, whereas Table 6 on p. 25, lines 423–424, sums to 912. Please state whether each value counts samples, stations, casts, profiles, or annual means, and document the population and exclusion rule used for every figure, table, and regression. Different values may be retained where scientifically justified, but a concise crosswalk should explain the counting unit, source population, exclusions, and resulting N for each output.
6. Interpretation relative to ESSD scope. The extent of scientific interpretation in Sects. 4–9 should be reduced or more clearly separated from the data-description purpose of the manuscript (pp. 16–27, lines 276–468). These sections include extensive trend, water-mass, apparent-age, and Atlantification interpretation. Figure 3 and the associated discussion on pp. 17–19, lines 284–316, are particularly close to a separate hypothesis-driven analysis. ESSD data descriptions should primarily establish data quality, uncertainty, plausibility, accessibility, and reuse. Please retain only the analyses needed to demonstrate those functions and move the fuller interpretation to the companion paper. If the Figure 3 regressions remain as limited examples of reuse, the existing caption already identifies the principal subsets, exclusions, and cast-mean unit of analysis. Please add reproducible input provenance and uncertainty for the reported slopes. Alternatively, the figure could be simplified to a descriptive example. I am not requesting additional hypothesis testing.
Minor comments
1. Dataset missingness. Please explain the meaning of blank values in the public dataset (pp. 9–10, lines 151–164, and New_Mirai_NABOS_DIC14C_dataset.xlsx, sheet DIC_isotope_data, Excel rows 2–256). Blanks occur in fields including Sample_ID, Depth_QC, Water_mass_initial, pMC, uncertainty, and DIC concentration. A blank may mean not measured, not recovered, not applicable, or not assessed, and these states have different implications for reuse and quality screening. Please define the applicable missing-data meanings and provide a concise completeness summary by variable and preferably by cruise. Missing legacy measurements do not need to be reconstructed.
2. QC status and narrative notes. I recommend separating assessment status from narrative explanation in column Depth_QC of New_Mirai_NABOS_DIC14C_dataset.xlsx, sheet DIC_isotope_data, Excel rows 2–256 (see also p. 9, lines 151–157). This field is blank for 187 of 255 records and otherwise mixes OK with narrative comments, which prevents reliable automated filtering. A controlled status such as assessed, not assessed, or flagged, accompanied by a separate note field, would be clearer. A community-standard flag should be used only if it maps honestly to the surviving legacy information.
3. Ambiguous date. Please clarify the date assigned to Excel row 129 in sheet DIC_isotope_data of New_Mirai_NABOS_DIC14C_dataset.xlsx (see also p. 9, lines 151–157, and Table 7 on p. 27, lines 462–468). In the same workbook, sheet Variable_dictionary, Excel row 8 defines Date_UTC using the YYYY-MM-DD format, whereas the record in DIC_isotope_data, Excel row 129 is given as 2006-09-02/03. This value could represent a date interval, an uncertain date, or a difference between UTC and ship time, and it is therefore neither machine-readable nor unambiguous under the current definition. Please encode a valid date or interval and document the applicable time basis. Separate start and end dates, or a single date accompanied by an uncertainty note, would both be acceptable.
4. Water-mass vocabulary. Please define and standardise the allowed values in column Water_mass_initial of New_Mirai_NABOS_DIC14C_dataset.xlsx, sheet DIC_isotope_data (see also p. 9, lines 151–157). This sheet uses Intermediate / deep transition (800–1500 m) in Excel rows 16–18, 37–38, 60–63, 73–75, 112–113, 158–160, and 252–253, and Intermediate/deep transition (800–1500 m) in Excel rows 141–143 and 190–191. Equivalent labels with different spacing or punctuation will be treated as separate categories by common analysis software. Please list the allowed values in sheet Variable_dictionary, Excel row 14, and use one harmonised form across the workbook, README, and repository metadata. If preserving the source wording is important, the original label could be retained in a separate source-label field.
5. DOI and non-DOI source links. Please distinguish persistent DOI links from general source URLs in New_Mirai_NABOS_DIC14C_dataset.xlsx, sheet Cruise_sources, Excel row 8 (see also p. 10, lines 158–160). The NABOS expedition report URL is stored in DOI_URL, although it is not a DOI. This may cause users or software to interpret the value as a persistent identifier. The field could be generalised, or DOI_URL could be left blank for NABOS while the report is retained in a source-URL field. No new identifier is required.
6. Apparent ¹⁴C-age calculation. Please define the apparent ¹⁴C-age calculation used in Sect. 7, Table 5, and Figure 4 (pp. 22–24, lines 381–417). The depth ranges are listed in Table 5, but the manuscript does not provide the apparent-age equation, reference year, decay convention, or uncertainty calculation. Without this information, readers cannot reproduce the reported ages or distinguish an apparent radiocarbon-equivalent age from a physical ventilation age. Please provide the calculation and state explicitly that the result is not a model-independent ventilation age. If the ages are intended only as descriptive indices, a concise equation and statement of assumptions would be sufficient.
7. Code availability and end-matter declarations. Please add the missing availability and declaration sections. Data availability ends on pp. 28–29, lines 469–492, and is followed directly by Conclusions, while Acknowledgements begin on p. 30, line 525. A Code availability section should identify the code or workflow used for the retained synthesis figures, tables, and regressions. Author contribution and Competing interests should also be added before Acknowledgements in the order required by the current Copernicus guidance. These may be concise statements. If some code cannot be deposited, please state what is available, how it can be accessed, and which outputs cannot be independently reproduced.
8. Dataset DOI in the Abstract. For the final accepted version, please add the dataset DOI and an in-text data citation to the Abstract (p. 1, lines 16–23). The current preprint already provides the functional DOI in the Data availability section and the reference list, so this is not a concern about access during review. However, ESSD manuscript guidance states that the Abstract of the final accepted article must include the functional dataset DOI and its in-text citation. A brief data citation accompanied by https://doi.org/10.17592/001.2026062501 would satisfy this requirement, with the exact placement following the journal’s preferred style.
Technical corrections1. Table 2 header layout. On p. 14, lines 236–258, prevent Halocline from breaking within the word. A deliberate line break before the depth range would be acceptable.
2. Figure text legibility. The smallest legend and note text in Figure 3 on p. 17, line marker 284, and Figure 4 on p. 23, line marker 388, appears small relative to the main figure labels. Please consider enlarging this text or moving detailed method information to the captions. It would also be helpful to confirm that the plotted groups remain distinguishable in colour and grayscale at final publication size.
3. Ascending numerical range. On p. 26, lines 448–450, report the range in ascending order as −9.0 to +5.2‰.Citation: https://doi.org/10.5194/essd-2026-493-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 226 | 80 | 39 | 345 | 43 | 37 | 38 |
- HTML: 226
- PDF: 80
- XML: 39
- Total: 345
- Supplement: 43
- BibTeX: 37
- EndNote: 38
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
My comment concerns the reference to the study by Payne et al. (2024) given in the manuscript. This study is listed in the references with the correct title, journal and doi, but the list of co-authors is not correct. More specifically, two people are listed who are not co-authors of this paper.
I cannot really understand how a mistake like this can happen, given that the doi of the paper is correct. My only explanation is the use of AI to generate this reference. I strongly encourage the authors to use suitable reference management software and to check the list of references carefully.