The PAGES CoralHydro2k Seawater &delta;<sup>18</sup>O Database: A FAIR-aligned compilation of seawater &delta;<sup>18</sup>O data to uncover 'hidden' insights from the global ocean

Atwood, Alyssa R.; Moore, Andrea L.; DeLong, Kristine L.; Long, Sylvia E.; Sanchez, Sara C.; Hargreaves, Jessica A.; Morris, Chandler A.; Pauly, Raquel E.; Dassie, Emilie P.; Felis, Thomas; Voelker, Antje H. L.; Murty, Sujata A.; Cobb, Kim M.

doi:10.5194/essd-2025-467

Preprints

https://doi.org/10.5194/essd-2025-467

Preprints

27 Aug 2025

| 27 Aug 2025

Status: this preprint is currently under review for the journal ESSD.

The PAGES CoralHydro2k Seawater δ¹⁸O Database: A FAIR-aligned compilation of seawater δ¹⁸O data to uncover 'hidden' insights from the global ocean

Alyssa R. Atwood, Andrea L. Moore, Kristine L. DeLong, Sylvia E. Long, Sara C. Sanchez, Jessica A. Hargreaves, Chandler A. Morris, Raquel E. Pauly, Emilie P. Dassie, Thomas Felis, Antje H. L. Voelker, Sujata A. Murty, and Kim M. Cobb

Abstract. The stable isotope values of seawater (δ¹⁸O and δ²H) provide valuable information on the exchange of water between the ocean, atmosphere, and cryosphere and on ocean mixing processes. As such, observational seawater δ¹⁸O and δ²H data place powerful constraints on hydrologic changes in the modern ocean. Seawater δ¹⁸O data are also essential for calibrating paleoclimate proxies based on the δ¹⁸O of marine carbonates and are an increasingly critical diagnostic tool for assessing model performance and skill in isotope-enabled global climate models. Despite their broad value, no centralized and actively-curated database for this type of data exists, even though a growing number of new seawater δ¹⁸O datasets have been generated over the last decade. As such, many seawater δ¹⁸O datasets remain ‘hidden’. To improve the accessibility of seawater δ¹⁸O data for the Earth Science research community, the Past Global Changes (PAGES) CoralHydro2k project has created a new, machine-readable, and metadata-rich database of observational seawater δ¹⁸O data, paired with seawater δ²H and salinity data, that is compliant with findability, accessibility, interoperability, and reusability (FAIR) standards for digital assets. The data has been collected from public databases and repositories, direct researcher data submissions, scientific papers, and student theses. In total, the PAGES CoralHydro2k Seawater δ¹⁸O Database contains over 18,600 data points with extensive metadata that makes the database suitable for a myriad of research applications. For hidden data, we searched for and included all datasets within the global ocean. For public data, our data collation efforts were focused on the upper 50 m from 35° N to 35° S (to aid in CoralHydro2k’s seawater δ¹⁸O reconstruction studies using δ¹⁸O and Sr/Ca in tropical-subtropical coral skeletons). We also provide a set of best practices to the community for reporting seawater isotope data in the future.

Received: 14 Aug 2025 – Discussion started: 27 Aug 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Status: final response (author comments only)

RC1:
'Comment on essd-2025-467', Anonymous Referee #1, 29 Sep 2025
This article presents a new machine readable metadata-rich database of observational seawater δ18O data, paired with seawater δ2H and salinity data, that is compliant with the FAIR standards. This is an important step towards improving the accessibility of seawater δ18O data for the Earth Science research community.
The article thus deserves publication. However, in its present form, the article is very short and contains many repetitions, so a number of points detailed below should be improved before it can be accepted for publication.

Main comments
The abstract mentions that the article provides “a set of best practices to the community for reporting seawater isotope data in the future”. However, these best practices are not described in the article. There is only one line (l. 419) mentioning that “the metadata template provides a set of best practices for reporting seawater isotope data in future studies”. This is not sufficient. The best practices the authors have in mind should be clearly communicated to the community in a dedicated section of the article.

The article should be reorganized to avoid repetitions. For instance there are many repetitions between the introduction, sections 2.1 and 5.3. Also, the statement that “the seawater δ18O data is paired with seawater δ2H, salinity, and temperature data, where available” is repeated in section 2.2. Please suppress as many repetitions as possible to improve the readability of the article.

It is unclear why the authors searched for datasets spanning all depths and latitudes for hidden data (l. 183) whereas they focused on the upper 50 m between 35°S and 35°N for published datasets. This seems inconsistent. Please clarify the rationale behind this approach. Also, the sentence l. 244 “Because the search for hidden datasets focused on the region between 35°N and 35°S […]” is in contradiction with the definition of the hidden data domain given l. 183, please clarify.

In the definition of level 6 metadata (l. 217-219), one does not see the difference between level 6 and level 5 metadata. I am guessing that level 5 metadata refer to water d18O, whereas level 6 metadata refer to secondary variables, like temperature. If so, this should be made clear in the definition of the different metadata levels in section 2.3.

294: another possible reason for the differences between model outputs and observations, is the local influence of E, P and runoff near the selected sampling sites compared to the open ocean far from islands (i.e. the ‘island’ or ‘continental’ effects) that is not well accounted for by the models due to their limited spatial resolution.

328-338: all these examples of applications assume that the salinity vs d18Osw relationship is stable in time, but it has been shown that this assumption is not verified in many instances: e.g. this relationship varies between monsoon and non-monsoon seasons (McConnell et al., 2009; Gosh et al., 2013), or in regions affected by sea ice formation and melting (Strain and Tan, 1993), and more generally in case of ocean circulation changes (Rohling and Bigg, 1998). This limitation should be explicitly stated.

411: The term “degree of freedom" is misused: In statistics, the degree of freedom refers to the number of random variables that cannot be determined or fixed by an equation. The way it is used here seems to imply that a higher degree of freedom would help define/understand the studied system. Please correct the sentence.

The last part of the sentence “the database can be used to better constrain the relationship between δ18Osw and salinity in the global ocean, and assess how this relationship varies in space and time"(l. 423) is an overstatement as far as paleoceanographic time scales are concerned. It is only true for seasonal variations. Please make this clear.

More minor comments
43-45: maybe mention that the stable isotopes of water composition is nearly conservative in sea water, when no phase exchange is involved (except for very small contributions related to chemical reactions, mostly on inorganic carbon, silicate, and nitrate cycles).

53-55: the link to marine biominerals and lipids is a bit indirect. It is important to mention, but it would be more appropriate to include it in the next paragraph (similarities with what is needed for paleoclimatic reconstructions).

82: in ‘addition to in situ atmospheric…’, mention ‘in situ oceanic…’

90-91: for precipitation, evaporation and salinity, there exists a very well structured international infrastructure related to GOOS (in addition to Argo, the asset of TAO and other tropical moorings, drifters, ship-of-opportunity measurements), please modify the sentence accordingly.

91: replace “select” by “selected”.

231-232: the correction for minor evaporation adjustment applied to some data points of the LOCEAN database is defined in the paper accompanying that database (Reverdin et al., 2022): “When breathing was not too large (resulting in an increase of less than +0.11‰ in d18O), we used the deviation from the expected d-excess relationship to S to estimate an adjusted d18O and dD (Benetti et al., 2017).” Note that this correction method is described in Appendix B of Benetti et al. (2017).

270-271: Fig. 4A is invoked to support the statement “only 13% of locations contain at least 12 measurements spanning two years within a 2° latitude x 2° longitude grid box”. However, Fig. 4A only shows the number of observations per 2°x2° grid cell, independently of their timing. Please clarify.

3C: the distribution of paired data in the southwestern Indian Ocean seems incomplete: most of the LOCEAN dataset in that region consists in data that include both d18O and d2H (as well as T and S) and cover the period 2008-2024.

Subtitle 4.1 should be removed because there is no section 4.2. One option could be to replace "Usage notes" by "Usage notes: General applications".

373-374 and 406-407: what do “researchers” refer to? Are these researchers who wish to add data to the CoralHydro2k database? or researchers who wish to download data from the CoralHydro2k database?

383: "scroll to the bottom of the webpage above and click on “Download Template”". It does not seem necessary to go into such a level of details in the article.

450: it seems awkward to provide the website of one laboratory and not of the others. Specifying the city and country or just the country could be enough for all the cited laboratories.

453: what is the link between the GEOTRACES 2021 Intermediate Data Product version 2 (IDP2021v2) and the CoralHydro2k Seawater δ18O Database? Please explain.

References
Benetti, M., Reverdin, G., Lique, C., Yashayaev, I., Holliday, N.P., Tynan, E., Torres-Valdes, S., Lherminier, P., Tréguer, P., Sarthou, G., 2017. Composition of freshwater in the spring of 2014 on the southern Labrador shelf and slope. J. Geophys. Res. Oceans 122, 1102–1121. https://doi.org/10.1002/2016JC012244
Ghosh, Prosenjit, Ramananda Chakrabarti, and S. K. Bhattacharya. "Short-and long-term temporal variations in salinity and the oxygen, carbon and hydrogen isotopic compositions of the Hooghly Estuary water, India." Chemical Geology 335 (2013): 118-127.
McConnell, Martha C., et al. "Seasonal variability in the salinity and oxygen isotopic composition of seawater from the Cariaco Basin, Venezuela: Implications for paleosalinity reconstructions." Geochemistry, Geophysics, Geosystems 10.6 (2009).
Rohling, Eelco J., and Grant R. Bigg. "Paleosalinity and δ18O: a critical assessment." Journal of Geophysical Research: Oceans 103.C1 (1998): 1307-1318.
Strain, Peter M., and Francis C. Tan. "Seasonal evolution of oxygen isotope‐salinity relationships in high‐latitude surface waters." Journal of Geophysical Research: Oceans 98.C8 (1993): 14589-14598.
Citation: https://doi.org/10.5194/essd-2025-467-RC1
RC2: 'Comment on essd-2025-467', Alessio Rovere, 30 Oct 2025

Dear Editor,
I have now completed my assessment of the manuscript under consideration.
The PAGES CoralHydro2k Seawater δ¹⁸O Database represents a new and comprehensive compilation of seawater stable isotope (δ¹⁸O and δ²H) and salinity data, developed in accordance with the FAIR principles (Findable, Accessible, Interoperable, Reusable). The initiative successfully addresses the lack of a centralized archive and makes previously “hidden” or scattered data accessible—information that is essential to the Earth Science research community. Its main goal is to support the calibration of coral-derived paleoclimate proxies and to enhance understanding of tropical hydrological processes and the performance of isotope-enabled climate models. The database, which includes more than 18,600 measurements collected between 1972 and 2021, constitutes the most extensive synthesis of marine isotope observations currently available. However, the authors rightly note the need for further globally coordinated sampling efforts, as the spatial and temporal coverage of the data remains uneven and incomplete.
Overall, the topic is valuable and worthy of publication. The dataset is impressive, and the manuscript reads well. Nevertheless, I believe a few aspects could be strengthened.
First, the description of the data fields seems too focused on what has been standardized from previous works, rather than providing clear guidelines for how new data should be reported. The authors should make an additional effort to describe how new data submissions should be formatted—perhaps specifying character limits or input rules for certain fields (some entries in the current database appear excessively long).
Second, the database’s structure—as a single CSV file rather than a true relational database—is a limitation. This format constrains automated validation (e.g., checking date formats, field lengths, or required entries) and makes the file cumbersome to handle. I am not suggesting that the authors completely redesign the structure at this stage, but these issues should at least be discussed, particularly regarding data quality control and verification.
Concerning usability, I reviewed the GitHub repository containing the example code. Currently, it provides only basic spatial plotting examples with minimal commenting and no use of Markdown cells to explain the workflow—one of the key advantages of Jupyter notebooks. Enhancing this documentation with better-annotated examples and more diverse use cases would greatly benefit potential users.
To further improve usability and encourage community contributions, I suggest providing a CSV template for new data submissions, along with a validation script to ensure that all fields are completed correctly and all mandatory information is present. This would enhance both the long-term sustainability of the resource and its adoption by the broader community.
I hope these suggestions will help the authors strengthen their work.

Citation: https://doi.org/10.5194/essd-2025-467-RC2

Viewed

Total article views: 1,374 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
1,273	76	25	1,374	30	41

HTML: 1,273
PDF: 76
XML: 25
Total: 1,374
BibTeX: 30
EndNote: 41

Views and downloads (calculated since 27 Aug 2025)

Month	HTML	PDF	XML	Total
Aug 2025	188	12	2	202
Sep 2025	938	23	11	972
Oct 2025	111	28	10	149
Nov 2025	36	13	2	51

Cumulative views and downloads (calculated since 27 Aug 2025)

Month	HTML	PDF	XML	Total
Aug 2025	188	12	2	202
Sep 2025	938	23	11	972
Oct 2025	111	28	10	149
Nov 2025	36	13	2	51

Viewed (geographical distribution)

Total article views: 1,352 (including HTML, PDF, and XML) Thereof 1,352 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 18 Nov 2025

Short summary

The stable isotopic composition of seawater is a valuable tool for studying the global water cycle in the past, present, and future. However, an active repository dedicated to archiving this type of data has been lacking, and many datasets remain hidden from public view. We have created a new database of observational seawater isotope data that is rich in metadata, publicly accessible, and machine readable to increase its availability and usability for a variety of Earth Science applications.


Total:	0
HTML:	0
PDF:	0
XML:	0

The PAGES CoralHydro2k Seawater δ18O Database: A FAIR-aligned compilation of seawater δ18O data to uncover 'hidden' insights from the global ocean

Viewed

Viewed (geographical distribution)

The PAGES CoralHydro2k Seawater δ¹⁸O Database: A FAIR-aligned compilation of seawater δ¹⁸O data to uncover 'hidden' insights from the global ocean