the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Oct 2026
Deep-time Sediment Sulfur Isotopes (DSSI): a database of sedimentary sulfur isotopic records through Earth history
Abstract. Sedimentary sulfur isotopic records constrain reconstructions of seawater sulfate evolution, pyrite burial, microbial sulfur metabolisms, ocean redox, anoxia, and Earth-system feedbacks. These reconstructions require isotopic values to remain linked to sulfur phase, age, locality, and source context. Here, we present the Deep-time Sediment Sulfur Isotopes database (DSSI), a compilation of marine sedimentary sulfur isotopic records in which each isotopic value remains linked to its sulfur phase and associated geological context, spanning 4,061–0 Ma. The released v1.0 datasheet (Guo et al., 2026; https://doi.org/10.5281/zenodo.21675881) contains 78,070 rows and 136 public fields; 64,990 rows contain at least one numeric sulfur isotopic value. Sulfur isotopic values are stored in unified isotopic fields and linked to a hierarchical phase classification comprising sulfur category, sulfur component, and sulfur carrier. The release also preserves stratigraphic age, modern and reconstructed paleogeographic coordinates, lithology, depositional setting, analytical method, and auxiliary geochemical information, where available. This structure supports source-traceable spatial and temporal synthesis that distinguishes sulfur phases without separating isotopic values from the geological metadata needed to interpret them.
- Preprint
(6576 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 14 Dec 2026)
-
CC1: 'Potential problems linking to source publications in DSSI', Weathering Intensity Working Group Sedimentary Geochemistry and Paleoenvironments Project, 03 Nov 2026
reply
I was excited to see this DSSI preprint and data file posted on ESSD. What a great data resource for the community! In going through the data file, however, I was a little confused about how some of the data relate back to the source publications. The authors have tried to do a good job with tracing back data to the sources, which I agree is very important in geoinformatics and for reproducibility. In looking at my published papers, though, there appear to be a lot of error in relating data back to source publications. To be clear, I understand how hard it can be to do large data compilations and I appreciate all the hard work the authors have put in. And I don’t want to make the mistake of “because there are some errors in compiling my samples, the whole endeavor is wrong." My concern is that it is not just a few errors, but that most of my papers appear to have incorrect source citations associated with them, or are included twice, and this may be indicative of a larger problem throughout the dataset.-The "Sperling 2018" samples from the Road River Group are listed as sourced from this abstract:10.1130/abs/2018am-317367That is a GSA abstract on the Cambrian explosion. That abstract has no samples or data associated with it, and is not even related to the later Paleozoic Road River Group. This Road River Group data is also unpublished and is not actually related to any publication whatsoever. So it is confusing how the authors got these unpublished data, samples, and an age model from an abstract that is actually on the Cambrian? I do have later publications on the Road River Group, but not with the sulfur isotope data.-My 2013 EPSL paper (10.1016/j.epsl.2013.04.003) is in the compilation twice, once with ages around 860 Ma and once with ages that are younger (which are the ones I would support).-Samples from the Tonian Beck Springs and Callison Lake Dolomite are associated with the link to this 2015 Annual Reviews paper:10.1146/annurev-ecolsys-110512-135808However, that Annual Reviews paper publishes no geochemical data (it is on physiology and doesn’t even have a supplement). I have made pyrite sulfur isotope measurements on E1101 and F1011 samples, but again they are not published, and certainly not in the Annual Reviews paper, so again it is unclear where this data is coming from in general and why it is associated with this paper.-Some E1-NC174 core samples are reported as coming from Loydell et al. 2013 in Palaeo^3 (10.1016/j.palaeo.2013.05.027). However, that paper does not present any sulfur isotope data. The sulfur isotope data for those samples are published in Stockey et al., 2020 Nature Communications.-Some of the other E1-NC174 samples are cited as being from Lai et al. 2025 (the DM-SED compilation). However, that is a compilation, not a source publication, and it would probably be better to cite the original publications here instead of DM-SED.Again, I know how much hard work must have gone into such a compilation, and I congratulation the authors on getting this near completion. As I said, this will be a great resource. I do worry though that there is some broader error in linking the compiled data to the source publications given the number of my publications that appear wrong, and I hope this can help the authors identify and fix that error.ReplyCitation: https://doi.org/
10.5194/essd-2026-610-CC1 -
CC2: 'Reply on CC1', Erik Sperling, 03 Nov 2026
reply
To be clear, the CC1 comment was by Erik Sperling (Stanford University). I included a Working Group name on an abstract I submitted through this portal ages ago, and apparently the Working Group name associated with my email cannot be changed!
Citation: https://doi.org/10.5194/essd-2026-610-CC2 -
AC1: 'Reply on CC1', Huyue Song, 06 Nov 2026
reply
General comment
I was excited to see this DSSI preprint and data file posted on ESSD. What a great data resource for the community! In going through the data file, however, I was a little confused about how some of the data relate back to the source publications. The authors have tried to do a good job with tracing back data to the sources, which I agree is very important in geoinformatics and for reproducibility. In looking at my published papers, though, there appear to be a lot of error in relating data back to source publications. To be clear, I understand how hard it can be to do large data compilations and I appreciate all the hard work the authors have put in. And I don’t want to make the mistake of “because there are some errors in compiling my samples, the whole endeavor is wrong." My concern is that it is not just a few errors, but that most of my papers appear to have incorrect source citations associated with them, or are included twice, and this may be indicative of a larger problem throughout the dataset.
Response:
We thank Erik Sperling for his positive assessment of DSSI and for identifying these problems. Our checks of the examples below show that some citations were transferred from earlier databases without being checked against the original publications. We also identified errors in our own assignment of references and ages. We will check the source of each dataset, paying particular attention to records obtained through other compilations, and correct the affected entries before the revised release.
The counts below refer to the submitted DSSI v1.0 release (78,070 records; https://doi.org/10.5281/zenodo.21675881). The corrections described here are intended for the revised release.
Comment 1
-The "Sperling 2018" samples from the Road River Group are listed as sourced from this abstract:
10.1130/abs/2018am-317367
That is a GSA abstract on the Cambrian explosion. That abstract has no samples or data associated with it, and is not even related to the later Paleozoic Road River Group. This Road River Group data is also unpublished and is not actually related to any publication whatsoever. So it is confusing how the authors got these unpublished data, samples, and an age model from an abstract that is actually on the Cambrian? I do have later publications on the Road River Group, but not with the sulfur isotope data.
Response:
Thank you for pointing out this problem. The measurements, sample information and ages were obtained from the Deep-Time Marine Sedimentary Element Database (DM-SED; Lai et al., 2025). We simply adopted the references provided in DM-SED without carefully checking the original publications. As a result, the GSA abstract was incorrectly cited as the source of these sulfur measurements.
We will further verify the original sources and publication status of all records obtained from other databases. All unpublished data obtained from these databases, including the Road River sulfur measurements you identified, will be removed from the revised DSSI database and associated analyses.
Comment 2
-My 2013 EPSL paper (10.1016/j.epsl.2013.04.003) is in the compilation twice, once with ages around 860 Ma and once with ages that are younger (which are the ones I would support).
Response:
After carefully reviewing the relevant records, we found that the sulfur isotope data from the 2013 EPSL paper had not been included twice. The two Excel supplementary files of this paper together contain 236 distinct section–height records, including 142 with sulfur isotope values and 94 with auxiliary data only. The submitted DSSI database contains the same 236 records and 142 sulfur isotope values. We also found that the author column was inconsistently formatted as “Sperling” and “Sperling et al.” This inconsistency may have given the impression that the same dataset had been included twice. We will correct these entries and standardize the author information throughout the database.
The age-assignment approach is described in Section 2.1 of our manuscript. The two age groups resulted from combining ages assigned during DSSI preparation with ages inherited from earlier compilations. Following your clarification, we will revise the affected records using the younger chronological framework you support and check other records for similar inconsistencies.
Comment 3
-Samples from the Tonian Beck Springs and Callison Lake Dolomite are associated with the link to this 2015 Annual Reviews paper:
10.1146/annurev-ecolsys-110512-135808
However, that Annual Reviews paper publishes no geochemical data (it is on physiology and doesn’t even have a supplement). I have made pyrite sulfur isotope measurements on E1101 and F1011 samples, but again they are not published, and certainly not in the Annual Reviews paper, so again it is unclear where this data is coming from in general and why it is associated with this paper.
Response:
After checking the relevant records, we found that the incorrect reference to the Annual Review paper was already attached to these sulfur isotope records in DM-SED (Lai et al., 2025). We imported the records and retained that reference without carefully verifying whether the review reported the measurements.
The submitted DSSI database contains 28 sulfur isotope records associated with this review. Based on your clarification, we will remove these 28 records from the revised database and associated analyses. We will also further verify all records obtained from other databases, correct inappropriate references and remove unpublished data from the revised DSSI database.
Comment 4
-Some E1-NC174 core samples are reported as coming from Loydell et al. 2013 in Palaeo^3 (10.1016/j.palaeo.2013.05.027). However, that paper does not present any sulfur isotope data. The sulfur isotope data for those samples are published in Stockey et al., 2020 Nature Communications.
Response:
Thank you for pointing out the correct source. After checking the original publications and supplementary data, we confirmed that the correct source of these 41 E1-NC174 sulfur records is Stockey et al. (2020), https://doi.org/10.1038/s41467-020-15400-y, Supplementary Data 1. We will correct the sulfur references and retain Loydell et al. (2013) as the source of the carbon data it reports.
Comment 5
-Some of the other E1-NC174 samples are cited as being from Lai et al. 2025 (the DM-SED compilation). However, that is a compilation, not a source publication, and it would probably be better to cite the original publications here instead of DM-SED.
Response:
After checking the relevant records against the original supplementary data, we confirmed that these 17 additional E1-NC174 sulfur isotope records were published by Stockey et al. (2020). We obtained them through DM-SED and cited the compilation without verifying the original measurement source. We will cite Stockey et al. (2020), https://doi.org/10.1038/s41467-020-15400-y, as the original sulfur source.
Concluding comment
Again, I know how much hard work must have gone into such a compilation, and I congratulation the authors on getting this near completion. As I said, this will be a great resource. I do worry though that there is some broader error in linking the compiled data to the source publications given the number of my publications that appear wrong, and I hope this can help the authors identify and fix that error.
Response:
Thank you very much for these important comments. We will give priority to checking references transferred from other databases and tracing the measurements to their original data tables. We will extend these checks to all remaining data sources and their sample identities, ages and associated measurements. Unpublished measurements will be excluded from the revised release.
After completing and verifying the corrections, we will update the affected statistics and figures, revise the manuscript accordingly, and release a new database version documenting the changes.
Citation: https://doi.org/10.5194/essd-2026-610-AC1 -
CC3: 'Reply on AC1', Erik Sperling, 07 Nov 2026
reply
I appreciate that the authors have done a nice job in checking with these issues. Thank you.
I guess the still-remaining issue in my head is that noting that these samples/data were brought over from Lai et al. 2025 (DM-SED), as stated in the response, doesn’t really solve the problem of how Lai et al. 2025 obtained data and age models for Road River Group samples from an abstract on a completely different topic (Sperling and Stockey, 2018, on the Cambrian radiation), or how they obtained unpublished pyrite sulfur isotope data from a paper that is focused on physiology, with no geochemical data at all. None of that makes any sense. I do acknowledge that I am part of the problem here because I was a co-author on Lai et al., but did not recognize these issues at the time.
In looking closer at the data, I wonder if a lot of the data in DM-SED that is listed as “new compilation” from the literature was actually just downloaded from the SGP database. The age models seem the same as SGP, which is not really what one would expect if people were actually developing age models de novo from the literature as a new data compilation. Or at least that is a striking coincidence! Also, the SGP database does include the unpublished sulfur isotope data in question. It also includes other unpublished sulfur isotope data that is included in DSSI, for instance the values from Canfield et al. 2008 in Science, that appear in DSSI associated with that publication but are also not listed in the supplemental data file from that paper. Those Canfield data were provided to SGP though and do appear in that database. So it would make more sense if much of these data incorporated into DM-SED were actually just taken from SGP instead of newly compiled from the literature. That would explain the overlap of the age models and the inclusion of unpublished data that are only held in SGP.
The authors are free, of course, to do whatever they think is best with this compilation. However, this looks like a really nice and useful compilation for the field and I would hate to see data excluded. Why not include as much sulfur isotope data as possible? That will be more useful to the field. My suggestion would be to figure out what was incorporated into DM-SED from SGP (if that is indeed the source of these data, although that seems pretty clear to me given that unpublished data was not obtained from a paper on physiology…), and then what was brought over to DSSI, and then simply properly credit SGP for those samples/data. This also shields DSSI from any errors in SGP. For instance, in researching the provenance of these data, I realized that SGP had incorrectly attributed pyrite sulfur isotope data from Johnston et al. 2013 Chemical Geology to a companion paper by Macdonald et al. 2013 Chemical Geology. By citing the correct source of your data, you are not responsible for that provenance error, it lies with SGP.
In any case, I wanted to reiterate that my comments here are aiming to help with the provenance of data and the tracing of data and to make this compilation better. I appreciate what these authors have done with these geochemical compilations and I look forward to seeing it published. Sincerely,
Erik Sperling
Stanford University
Citation: https://doi.org/10.5194/essd-2026-610-CC3 -
AC2: 'Reply on CC3', Huyue Song, 11 Nov 2026
reply
Thank you for your thoughtful follow-up and helpful suggestions. We will trace the relevant measurements and age models through DM-SED and SGP, verify their original sources, and correct the source references in DSSI. We will compare sample identifiers, sulfur isotope values and age information to establish which records were transferred from SGP. Where SGP is confirmed as the source, we will credit it explicitly and document the transfer through DM-SED. We will also distinguish the sources of the measurements from those of the age assignments, so that both can be traced clearly.
We will retain all data confirmed to have been published and continue to collect additional published sulfur isotope measurements to make the compilation as comprehensive as possible. Their sources and associated uncertainties will be clearly documented in the revised database.
We also plan to maintain and update DSSI beyond the present release, with updates approximately every 1-2 years to correct identified errors and incorporate newly published data. These updates will support continued use of the database by both our team and the wider community in future research. We appreciate your continued help in making DSSI a more complete and reliable resource for the community.
Citation: https://doi.org/10.5194/essd-2026-610-AC2
-
AC2: 'Reply on CC3', Huyue Song, 11 Nov 2026
reply
-
CC3: 'Reply on AC1', Erik Sperling, 07 Nov 2026
reply
-
CC2: 'Reply on CC1', Erik Sperling, 03 Nov 2026
reply
-
RC1: 'Comment on essd-2026-610', Stephen J. Mojzsis, 09 Nov 2026
reply
Dear Colleagues, thank you indeed for this very useful work that will be an important addition to the biogeodynamic analysis of the evolution of S isotopes in the Earth system. I echo the previous comments, and wish to add a few of my own at this time.
The assignment of 4061 Ma in the ages of Uveges et al. 2023 is problematic. None of the rocks of the Nuvvuagittuq Supracrustal Belt are so old; what the authors quote are model ages from Sm/Nd errorchrons. Instead, the minimum ages of these rocks is about 3800 Ma as pointed out in a number of papers.
Unfortunately, the papers by Jonathan O'Neil continue to make unfounded claims. It does not change the nature of the present compilation to put a necessary "FLAG" on these data for ages, but to accept at face value the ca. 4061 Ma age assignment is an error and propagates a falsehood.
- Steve Mojzsis, Bayreuth BGI
Citation: https://doi.org/10.5194/essd-2026-610-RC1 -
AC3: 'Reply on RC1', Huyue Song, 11 Nov 2026
reply
Thank you for highlighting this important chronological issue. We agree that 4061 Ma should not be treated as an established formation age for these samples. Following your recommendation, we will adopt approximately 3800 Ma as a conservative reference age for the affected records. We will check the original chronological studies and their applicability to the relevant samples when implementing this revision, and update the manuscript and associated analyses accordingly.
Citation: https://doi.org/10.5194/essd-2026-610-AC3
-
AC3: 'Reply on RC1', Huyue Song, 11 Nov 2026
reply
Data sets
Deep-time Sediment Sulfur Isotopes (DSSI): a database of sedimentary sulfur isotopic records through Earth history W. Guo et al. https://doi.org/10.5281/zenodo.21675881
Model code and software
DSSI Reproducibility Package: Figure Generation and Paleogeographic Coordinate Reconstruction W. Guo https://doi.org/10.5281/zenodo.21666567
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 425 | 147 | 34 | 606 | 36 | 30 |
- HTML: 425
- PDF: 147
- XML: 34
- Total: 606
- BibTeX: 36
- EndNote: 30
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
