the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Jul 2025
The global database of deep-time marine nitrogen isotope data
Abstract. Stable nitrogen isotope records preserved in marine sediments provide critical insights into Earth’s climate history and biospheric evolution. Although numerous studies have documented nitrogen isotope (δ15N) records across varied geological ages (Archean to Recent) and paleogeographic settings, the scientific community remains constrained by the absence of a standardized database to systematically investigate their spatiotemporal evolution. Here, we present the database of Deep-time Sediment Nitrogen Isotopes in Marine Systems (DSMS-NI), a comprehensive global compilation of δ15N data and associated geochemical parameters, spanning a vast collection of sediment samples dating from the Recent to the Archean. This database encompasses 71 040 δ15N records derived from 424 publications, systematically organized with 29 metadata fields categories (e.g., chronostratigraphic ages, coordinates, lithology, metamorphic grade, sedimentary facies, references) encompassing 1 927 829 metadata. This repository further incorporates 130 proxy data fields, including 285 715 geochemical data spanning total organic carbon (TOC), total nitrogen (TN), and organic carbon isotopes (δ13Corg), major and trace elements and iron species. These integrated parameters enable evaluation of sample fidelity and factors influencing δ15N signatures. The DSMS-NI database will facilitate research across key geological intervals such as the Permian-Triassic boundary and the Cretaceous ocean anoxic events. Researchers can leverage temporal and paleogeographic information, alongside geochemical data, to conduct spatiotemporal analyses, thereby uncovering changes in deep-time marine nitrogen cycles and paleoenvironmental conditions. The database is open-access via the Geobiology portal (https://geobiologydata.cug.edu.cn/, last access: 30 April 2025), allowing users to access data and submit new entries to ensure continuous updates and expansion. This resource represents a vital foundation for studies in paleoclimate, paleoenvironment, and geochemistry, offering essential data for understanding long-term Earth-system processes. The data files described in this paper are available at https://doi.org/10.5281/zenodo.15117375 (Du et al., 2025a).
- Preprint
(2867 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on essd-2025-377', Anonymous Referee #1, 11 Sep 2025
-
RC2: 'Comment on essd-2025-377', Anonymous Referee #2, 25 Sep 2025
General Comments
Du et al. have established the Database of Deep-time Sediment Nitrogen Isotopes in Marine Systems (DSMS-NI) in this work. By integrating previously published datasets and supplementing them with newly collected data, this dataset represents a comprehensive compilation in the field of marine sediment nitrogen isotopes. It includes 71,040 δ15N data points spanning from the Archean to the modern, and covers a wide range of sample types, from fossil materials to bulk sediments, kerogen, and others. The authors have also made substantial efforts to enrich each nitrogen isotope data point with additional metadata, such as modern latitude and longitude, paleolatitude and paleolongitude, age information, lithology, depositional water depth, and, where available, other geochemical information from the same stratigraphic horizon. This endeavor greatly expands the potential for future research based on this database.
At the same time, several issues should be addressed to make the dataset more robust and user-friendly:
1. Although the manuscript carefully explains the metadata fields, the database itself is not accompanied by a Data Descriptor File (ReadMe file). Providing such a file would substantially improve data usability and accessibility for the community.
2. In the current dataset, materials are categorized into sediments, foraminifera, diatoms, etc. However, foraminifera- and diatom-bound nitrogen isotope measurements are reported as bulk δ15N values, while porphyrin and kerogen are listed separately. This classification seems inconsistent and somewhat confusing. A clearer approach might be to keep “sediments” and sedimentary rock types as the main material categories, and then include foraminifera, diatoms, porphyrin, and kerogen as specific entries in the data fields.
3. While the dataset is comprehensive, it does not appear to include coral-bound δ15N records, which have become an increasingly important proxy for reconstructing Holocene nitrogen cycling. Incorporating, or at least acknowledging, this data type would further strengthen the completeness of the compilation.
4. The dataset would benefit from including a field specifying the analytical method used for δ15N measurements. Different techniques (e.g., EA flash combustion, denitrifier method, chemical oxidation, offline combustion) may introduce distinct biases, and distinguishing these would be important for future analyses.
5. Some citation and metadata errors are present. For example, data attributed to Alt-Epping U. et al. (2009) shows negative ages, and some entries cite conference abstracts (e.g., Wang X. et al. 2021, Goldschmidt 2021 abstracts) rather than the final peer-reviewed sources. In addition to automated quality control, it would be helpful to perform manual spot-checking of a representative subset of entries against the original literature. If the authors could demonstrate that the error rate is low, this would increase confidence in the dataset.
6. It would be helpful if the authors could clarify their plans for future updates and maintenance of DSMS-NI, and whether mechanisms will be provided for incorporating newly published data or community-submitted contributions. This would further increase the long-term value and sustainability of the dataset.
In addition, I agree with Anonymous Referee #1 that while the descriptive statistics on the dataset are useful, some of the interpretations presented in the manuscript may be too strong at this moment.
Overall, this manuscript makes an important contribution to the community, and I support its publication after the authors address the above points.Specific Comments
• Figure 4: Some plotted points appear to fall on land, especially for the Cretaceous interval; please check the basemap.
• Lines 374–376: This sentence lacks sufficient supporting evidence or a figure. Please clarify.
• Line 400: Quaternary sites also provide extensive latitudinal coverage.
• Line 457: After the phrase “we have also provided a software tool on Zenodo”, please add “see 8. Code availability”.Citation: https://doi.org/10.5194/essd-2025-377-RC2
Data sets
The global database of deep-time marine nitrogen isotope data Yong Du, Huyue Song, Thomas J. Algeo, Hui Zhang, Jianwei Peng, Yuyang Wu, Jiankang Lai, Xiang Shu, Hanchen Song, Lai Wei, Jincheng Zhang, Eva E. Stüeken, Stephen E. Grasby, Jacopo Dal Corso, Xiaokang Liu, Daoliang Chu, Li Tian, Qingzhong Liang, Xinchuan Li, Hong Yao, Haijun Song https://doi.org/10.5281/zenodo.15117375
Model code and software
Python code used to validate the dataset and make the figures Yong Du https://doi.org/10.5281/zenodo.15758073
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,333 | 65 | 40 | 1,438 | 39 | 55 |
- HTML: 1,333
- PDF: 65
- XML: 40
- Total: 1,438
- BibTeX: 39
- EndNote: 55
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
Du et al. have compiled an extensive database for marine nitrogen isotopes, covering data dated back to Archean. The database organizes data from over 400 publications and also makes use of existing geochemical databases. Building on existing published data compilations, this deep time marine nitrogen isotopes dataset covers a longer time span, centralizes a range of geochemical proxies and works to provide age estimates that are convenient to use. The authors assembled a total of 71040 δ15N data and 285715 geochemical data that allow quality evaluation and paleoenvironmental interpretations. This database is also accessible to all and the authors plan to continue improving and enlarging the repository upon future research.
This paper goes through data compilation, data summary and validation for data quality. The structure is organized and easy to follow. The authors put into effects to calibrate age estimates that can be used readily and also are in reference to an updated geochronological framework. However, they did not provide sufficient discussion on age uncertainties that their age adjustment might incur. Aside from accessibility of the dataset, the data quality and uncertainty are also essential. Similarly, the discussion on data quality check and validation can benefit from some additional details and explanation (see specific comments).
The other concern I have is the authors’ interpretation on database statistics. Although the key objective of this paper is not on paleoenvironmental reconstruction, the authors showcased some new insights the database can bring. I would suggest them to present these findings with interpretation that phrased as proposed or possible mechanism/idea, avoiding making strong arguments for rigor, for example, the variation of δ15N throughout time, and the latitudinal trend of δ15N. In the modern ocean, the spatial pattern of δ15N is not dominantly affected by latitudes. The apparent higher δ15N values at lower latitudes in Cenozoic might be driven by biased focus on regional oceanographic feature, such as the oxygen minimum zones in the tropical Pacific.
I think the presented database would be a beneficial addition for the community studying climate, marine environment and ecology in Earth history. After revisions and clarifying data uncertainties and quality validation, this database can motivate further investigations for understanding the marine environment and biogeochemical cycling throughout Earth history.
Specific comments
Line 85: suggest change to “survey of δ15N records on bulk sediments and biominerals deposited within”.
Line 96: In Table 1, it would be more clear if having “broad age”, “crude age” and “fine age” defined as table annotations or include the definitions in the main text.
Line 148: What types of settings are considered highly heterogeneous? It would be good to have a couple of examples. Citation or explanation is needed here.
Line 159: The citation “Farmer et al., 2021” is not listed in the reference. Also, more suitable citations for the denitrifier-based method are Ren et al., 2012 in Limnology and Oceanography, and Smart et al., 2018 in GCA.
Line 165: It is unclear what “repeated measurements of the same sample” are. Are these replications of the same sample, or measurements for the same type of samples?
Line 169: What are the criteria as reasonable explanations for extreme δ15N? Are the explanations based on the discussion in the source publications? Please include citation or example here.
Line 193: In Table 2, under field name “Isotopes”, change “the isotope composition” to “the isotopic composition”.
Line 193: In Table 2, under field name “Reference”, change the typo “formated” to “formatted”.
Lines 234-237: The explanation for age adjustment here seems to be repetitive as explained in the “Age model calibration” (Lines 319-322).
Line 308: An expected δ15N range needs to be defined here, as what δ15N values would be considered “unusually high or low”? If so, do the outlier δ15N get removed from the database? What validation process or criteria that determine whether the data is included or not?
Lines 319-322: Does the adjustment for outdated age estimations only refer to the boundaries between geologic stages of the International Chronostratigraphic Chart? Also, it is unclear that how the age-depth model is developed, assuming a constant sedimentation rate? Were any other geochronological tie points considered for the age-depth model, such as the paleomagnetic reversals? Adjusting the age with “the median age of the corresponding geologic interval” might introduce age uncertainties. I suggest the authors to add some discussion or comments on the age uncertainties, which would provide more rigorous quality evaluation for users in the future.
Line 367-368: The sentence “the Cenozoic has the highest overall peak” is unclear. Does it refer to the highest peak density? Or do you mean the Cenozoic has the highest mode δ15N value? The sentence “lower peaks in the Paleozoic and Mesozoic” has the same issue.
Lines 369-371: Does the peak δ15N occur in the Late Cretaceous or at the K/Pg boundary? Also, I did not observe any notable δ15N peaks in mid-Triassic, Jurassic, and Early Cretaceous from Fig. 7. The interpretation that δ15N values “align with greenhouse-icehouse climate cycles” might be too strong, since no correlation is shown for the analysis in the paper. Climate cycles imply that repeated up-and-down feature can be observed, whereas the δ15N record through Phanerozoic does not show clear cyclic pattern.
Lines 411-413: The Fig. 9 does not include data for the Ediacaran. The citation of “Fig. 9” should move to the first half sentence.
Database file (DSMS-NI_v0.2): Under the column for “Material”, a few options refer to the same thing – foraminifer, foraminifera, foraminifers, and planktic foraminifera. All of these data are foraminifera-bound organic matter nitrogen isotopes. Different taxa of planktic/planktonic foraminifera might be used for analysis in these source publications. It could be very useful to clarify the taxonomic information.