the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 May 2026
SISAL_monv1: a global database of cave monitoring observations
Abstract. Cave monitoring is the process of collecting observational data such as microclimate conditions, hydrogeochemistry and water movement in cave systems. The most common motivation is for speleothem science – the reconstruction of past climate and environmental variability from cave formations such as stalagmites. Applications also include the monitoring of potential recharge to aquifers, as well as cave conservation. PAGES-SISAL (Past Global Changes – Speleothem Isotope Synthesis and AnaLysis), an international working group focused on speleothem science, has created a new global database of cave monitoring data consisting of drip rates and drip water isotopes, and the modern carbonate precipitates that form from these drip waters (so called farmed carbonates, grown on artificial substrates). Moreover, we report meteoric precipitation amounts and water isotopes at or near to cave sites. The draft version of the database can be found at https://repo.researchdata.hu/privateurl.xhtml?token=43a43257-f06e-4dc5-9e96-6603eabe775f, which will be available under doi: 10.5158/ARP/29W5J3 upon publication. The database contains datasets from 75 caves, with summary information including meta-data on location, elevation, cave depth, lithology, measurement methods and citations for original publications. Speleothem records previously curated by SISAL in SISAL speleothem database versions and corresponding to monitored drip sites are also identified in the new SISAL monitoring database. To supplement observational data gaps and provide accessible and consistent climate data for users, surface climate (precipitation, evaporation, temperature) and meteoric water isotope data extracted from global climate model products are also included in the database.
- Preprint
(1571 KB) - Metadata XML
-
Supplement
(714 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2026-31', Anonymous Referee #1, 29 Jun 2026
-
AC1: 'Reply on RC1', Pauline Treble, 31 Aug 2026
We sincerely thank RC1 for their time reviewing our manuscript and for providing helpful feedback that we have adopted to improve the user interaction with the database and also for improving the manuscript. We provide point-by-point responses (in bold) below.
- The database includes site-specific precipitation amount and precipitation isotope sample data, but it also includes monthly precipitation and precipitation-isotope values extracted from gridded products such as MSWEP and IsoGSM. I suggest that the authors explain this relationship more explicitly. In particular, it would be helpful to clarify whether gridded values are extracted at the cave coordinates or precipitation-monitoring-site coordinates, how users should handle differences in temporal resolution between observed samples and monthly products, and whether monthly precipitation isotope values are precipitation-weighted means. This would help users avoid treating local observations and gridded estimates as interchangeable.
We thank RC1 for pointing out where we could make the relationship between the observed and product data in the SISAL_monv1 database more clear. The product data was extracted at the caves sites and we will clarify this in the revised manuscript by adding the following new sentence to Section 2.5 L278 ‘data were extracted using the cave site co-ordinates (‘site’ in Section 2.3)’. We appreciate the point that RC1 raises with regards to the differences in temporal resolution between observed and product data. The motivation for providing the precipitation isotope product data was to fill a data gap for cave sites where precipitation isotope measurements were not available (outlined in last paragraph Section 2.5) rather than to facilitate comparisons between observed and product data. We highlight that the temporal resolution is stated for both observed and product data in SISAL_monv1 and it is also stated L276 that the original temporal resolutions of the product data vary from hourly to daily but that we chose to provide the product data at monthly resolution as it most closely aligns with the temporal resolution of cave dripwater sampling. Table 2 provides information to assist users if they wish to consult the original global products should they require higher resolution data. Furthermore, we have now included the code that we used to extract the product data on the repository that users can access. Excellent point regarding whether the gridded precipitation isotope values are precipitation-weighted means. We can confirm that they are precipitation-weighted means and will update L275 “and represent precipitation-weighted means.” With regards to observed precipitation isotope data, we input data as supplied either by users or as could be found in published sources in its original format. No down-sampling of precipitation isotopes to monthly resolution was undertaken.
- It would be very useful for users to quickly know the duration of the monitoring data available for each site and each data type. At present, users may need to inspect the sample tables directly to determine the start and end dates of precipitation isotope monitoring, drip-water isotope monitoring, drip-rate monitoring, or modern carbonate sampling. I'm not sure if it is easy to add summary fields, either in the metadata tables or as an additional summary table, giving the monitoring start date, end date, duration, and possibly number of samples for each site and data type. This would greatly improve the discoverability and usability of the dataset, especially for users interested in comparing records across sites or selecting datasets suitable for time-series analysis.
Many thanks to RC1 for providing this advice to improve user interaction with the database. We agree that this suggestion would help and have updated the SISAL webApp to include an overview of the sampling duration on the summary table that appears for each site. This change to the webApp has been implemented and is accessible to the reviewers.
- Missing parentheses around Lines 363-364. Thank you, typo will be addressed.
Citation: https://doi.org/10.5194/essd-2026-31-AC1
-
AC1: 'Reply on RC1', Pauline Treble, 31 Aug 2026
-
RC2: 'Comment on essd-2026-31', Anonymous Referee #2, 30 Jul 2026
General Comments:
Most published speleothem studies rely on cave monitoring data from a single cave location to aid in the interpretation of speleothem paleoclimate records. The database presented in this work, SISAL_monv1, collates cave monitoring data associated with cave air, drip water, soil, meteoric precipitation, surface air, and farmed calcite from 147 global caves to allow for easy regional comparisons. The authors source their monitoring data from a combination of peer-reviewed published studies, theses, and a community-wide call. They supplement the monitoring data with climate and precipitation output from reanalysis products and isotope-enabled models. Though the primary application of this database is for speleothem paleoclimatology studies, researchers exploring karst, aquifer, and groundwater dynamics as well as proxy-model intercomparison will also benefit from this thorough and comprehensive database. The authors have taken great care to follow accessible data management practices, including offering a GUI webApp for data querying without knowledge of coding languages and providing extensive metadata. I strongly support publication of this manuscript following one general clarification and some minor specific comments as described below.
The authors report a total of 75 cave sites in the abstract. When querying the entire map in the SISAL webApp GUI for SISAL_monV1 using filter Type 2 (Lat from -90º to 90º, Lon from -180º to 180º), 75 cave sites appear, but only 72 of these have an associated site_id field and only 73 have associated coordinates. This is also inconsistent with Table S1, which reports a total of 147 sites. 103 of these sites have drip water data and 137 sites have cave air data. The authors should reconcile these discrepancies and provide a consistent definition of the size of their dataset. Further, it would be helpful to implement a standardized procedure to contribute to the database, where authors can upload published work or theses and the working group can periodically review, edit, and incorporate these new data. In the absence of this workflow, a dedicated data submission point of contact would encourage a more comprehensive database for future work.
Specific Comments:
Ln 69: Change correct to robust.
Ln 95: Specify modern drip water. Maybe: “modern drip water hydrochemistry, providing a modern context for speleothems.”
Ln 106: I am not convinced by the statement that “site specific models may not be useful or practical,” and would recommend a citation here if the authors wish to retain it. It seems to me that site-specific monitoring data should be used to investigate models of individual cave environments, since cave microclimate (and corresponding isotope fractionation) can vary substantially even between neighboring caves of similar lithology and depth or neighboring chambers within the same cave. This principle is mentioned in Ln 321 (“calibrating forward models…”) and is explored in mechanistic models such as CaveCalc (Owen et al., 2018).
Ln 108: I suggest adding an additional application of comparison to laboratory experiments (e.g., Dreybrodt et al., 2016, Wiedner et al., 2008).
Ln 111-114: The authors state the cave monitoring group site is no longer being updated and provide a last accessed date of 12-08-2019. Does this access date reflect the most up-to-date version of the website? If so, would be helpful to state.
Ln 128-129: “Drip water stable isotopes are the most commonly measured monitoring variable” – I recommend specifying that you are referring to this dataset compilation. Further, consider specifying “most commonly measured drip water monitoring variable,” as I would assume cave air temperature is more common (as substantiated by Table S1). See also Ln 301-302 and Figure 3 title “Monitoring data by region.”
Ln 204: What variable does timing of sample collection correspond to? Include in parentheses. See also Ln 206.
Fig 3: If there are no sites or data in the “Other” category, I recommend removing this from the graphs and legends. Consider also removing the other categories with no data (arid, hyper arid) and instead stating data paucity for those climate classifications in the figure caption.
Table S1 Row 2: Would be helpful to include units, unless they vary substantially across publications.
Table S1 Column BM: Would be helpful to separate time span and sampling resolution.
Technical Corrections:
Ln 137-138: Sentence is cut-off.
Ln 184: “Includes” should be lowercase to match formatting of other bullet points.
Ln 216-217: “In addition it provides information on which surface did the modern carbonate grow and what was its mineralogy” – this feels strangely worded. I recommend: “In addition it provides information on the surface substrate and mineralogy of modern carbonate growth.”
Ln 216-219: “Here modern carbonates refer to…” is repeated twice.
Ln 227: Last sentence (“The colors of the format of the particular field.”) is incomplete.
Ln 231: Missing parenthesis after Supplement.
Ln 363: Remove extra parenthesis before citation.
Table S1 Column H: “Humidity” misspelled.
Table S1 Column N: “Dripwater” should be two words here to be consistent with its other appearances in the paper.
Citation: https://doi.org/10.5194/essd-2026-31-RC2 -
AC2: 'Reply on RC2', Pauline Treble, 31 Aug 2026
We sincerely thank RC2 also for their time reviewing our manuscript and for providing helpful feedback, in particular for cross-checking the size of the dataset as it appears in the WebApp versus the manuscript. We have now addressed this and also updated the manuscript to indicate plans for future integration of new datasets into the database. We provide a detailed point-by-point response in bold below.
The authors report a total of 75 cave sites in the abstract. When querying the entire map in the SISAL webApp GUI for SISAL_monV1 using filter Type 2 (Lat from -90º to 90º, Lon from -180º to 180º), 75 cave sites appear, but only 72 of these have an associated site_id field and only 73 have associated coordinates. This is also inconsistent with Table S1, which reports a total of 147 sites. 103 of these sites have drip water data and 137 sites have cave air data. The authors should reconcile these discrepancies and provide a consistent definition of the size of their dataset. Further, it would be helpful to implement a standardized procedure to contribute to the database, where authors can upload published work or theses and the working group can periodically review, edit, and incorporate these new data. In the absence of this workflow, a dedicated data submission point of contact would encourage a more comprehensive database for future work.
We sincerely thank RC2 for their time spent cross-checking this detail and finding this error. We can confirm that there are 75 sites in the database and that each site has an associated ID field and coordinates. The missing ID and coordinates have now been fixed and appear for all 75 sites. Many thanks again for pointing this omission out.
With regards to Table S1, it indeed lists more sites than appear in SISAL_monv1. This Table summarises sites and their meta-data that appear on a website called “cave monitoring database” described in Section 2.1. However, this website is a database of meta-data only, it does not contain any actual data nor links to any data. It represents an early effort by the international speleothem science community to provide a website that summarises where research is being conducted, type of research and by whom. It was thus a very useful starting point for us to i) compile a list of the type of data the community were collecting in order to decide what could potentially be included in the SISAL_monv1 database (see Section 2.1); and ii) whom to contact to invite contributions of data to the SISAL_monv1. Our approach for obtaining the data is outlined in Section 2.2 and describes the effort we undertook to include data into SISAL_monv1. While the website lists monitoring data sites, some of these data have not been published and the contact person (if they could be contacted) may have declined the opportunity to include their unpublished data. In other cases, data may have been published but appeared in figure format only requiring the corresponding author to either supply the original data or consent to having the data digitised (see section 2.2). We were able to obtain many datasets this way but in some cases we received no response and in accordance with SISAL’s policies, we did not include digitised data without consent from the corresponding author. As section 2.2 details, we also put general callouts to our networks including the SISAL email list and the Karst Record Conference network for data contributions to raise awareness of the opportunity to have data included in SISAL_monv1.
Thank you also for raising the question of how to manage inclusion of data in the future. Over the past year, lead Stewards Laura Endres and Nikita Kaushal have been working with Neotoma developers to extend the Neotoma data model to support concepts in SISAL and then to vet and upload the SISAL 3.0 data to Neotoma, which resulted in 374 speleothem sites with over 900 speleothem datasets being added to Neotoma https://www.neotomadb.org/news-events/welcome-to-sisal-3-0 . In addition, a manuscript is currently being drafted for ESSD, that describes how we integrated SISAL 3.0 into the Neotoma Paleoecology Database, a “database of databases.” This move significantly reduces the human effort needed to manage SISAL in the future, while also allowing palaeoclimatologists and other researchers to access a broader range of related datasets in one place. There have not yet been efforts to incorporate SISAL_monv1 into Neotoma as the current focus is on using SISAL_v3 (speleothem records) as the test case. Following this the intention is that SISAL (monitoring) will be considered for this process.
Specific Comments:
Ln 69: Change correct to robust. Change will be made as suggested.
Ln 95: Specify modern drip water. Maybe: “modern drip water hydrochemistry, providing a modern context for speleothems.” Partial change will be adopted. Dot point will be re-written as follows: ‘Highlight the dominant controls on drip water hydrochemistry, providing a modern context for speleothems, and identify, for example, climatic forcing versus local factors.’ It did not seem necessary to include ‘modern’ in reference to the drip water.
Ln 106: I am not convinced by the statement that “site specific models may not be useful or practical,” and would recommend a citation here if the authors wish to retain it. It seems to me that site-specific monitoring data should be used to investigate models of individual cave environments, since cave microclimate (and corresponding isotope fractionation) can vary substantially even between neighboring caves of similar lithology and depth or neighboring chambers within the same cave. This principle is mentioned in Ln 321 (“calibrating forward models…”) and is explored in mechanistic models such as CaveCalc (Owen et al., 2018). We thank RC2 for the opportunity to improve the clarity of the sentence. We do think there has been a mis-communication here and acknowledge that perhaps the sentence was not clear on the point we wished to make. RC2 states “...site specific monitoring data should be used…” (see our underlining of above text) however, in this sentence we are referring to “site specific models” rather than “site specific monitoring data”. To clarify, we will changed the sentence from “This may be particularly important for global initiatives as site specific models may not be useful or practical” to “This may be particularly important for global initiatives as models built and parameterised for a specific site may not be useful or practical when applied at a global scale.”
Ln 108: I suggest adding an additional application of comparison to laboratory experiments (e.g., Dreybrodt et al., 2016, Wiedner et al., 2008). Thank you for the suggestion for this important point. We will add the additional point ‘Constrain laboratory experiments (e.g., Wiedner et al., 2008) and/or theoretical models of carbonate precipitation (e.g., Dreybrodt and Romanov, 2016) by providing farmed carbonate data grown under field conditions for comparison.’
Ln 111-114: The authors state the cave monitoring group site is no longer being updated and provide a last accessed date of 12-08-2019. Does this access date reflect the most up-to-date version of the website? If so, would be helpful to state. Thanks for the suggestion, we have communicated with the website owner and the website had been updated since we last accessed. For completeness, we updated the information in Table S1 to be consistent with the meta-data available on the website. The sentence will now read: “From this dataset (accessed 27/08/2026), we identified…”
Ln 128-129: “Drip water stable isotopes are the most commonly measured monitoring variable” – I recommend specifying that you are referring to this dataset compilation. Further, consider specifying “most commonly measured drip water monitoring variable,” as I would assume cave air temperature is more common (as substantiated by Table S1). See also Ln 301-302 and Figure 3 title “Monitoring data by region.” Thanks for the suggestion to clarify the sentence. We will implement the suggestions.
Ln 204: What variable does timing of sample collection correspond to? Include in parentheses. See also Ln 206. Thank you, we will insert text “(start and end date/time)” here and also in descriptions of precip_sample, drip_iso_sample, drip_rate_sample and mod_carb_sample.
Fig 3: If there are no sites or data in the “Other” category, I recommend removing this from the graphs and legends. Consider also removing the other categories with no data (arid, hyper arid) and instead stating data paucity for those climate classifications in the figure caption.
Thank you, we have identified the problem which was that some cave sites were missing from the global product data which produced categories that appeared empty which became labelled as ‘other’. We have now updated the site list used to extract the global product data to contain all cave sites and this will remove the ‘other’ category. We have also removed the arid and hyper-arid classes as suggested and will update the caption to say that no data exist from these classes.
Table S1 Row 2: Would be helpful to include units, unless they vary substantially across publications. Thanks for the suggestion. Table S1 tabulates meta-data from https://cavemonitoringgroup.wordpress.com/ (section 2.1). We can confirm that units are not supplied with the variables listed on the website and that Table S1 is supplied only to support our tally of sites and variables as background information for our study. It would take a considerable amount of our time to track down units (which could vary between sites) using publications and likely impossible to confirm for unpublished data while adding no real value to our study. No change made.
Table S1 Column BM: Would be helpful to separate time span and sampling resolution. We thank RC2 for this suggestion although again emphasis that the information in Table S1 is merely a tabulation of background meta-data. It was not used as input data for our study and is not required to be in machine-readable format. No change.
Technical Corrections:
Ln 137-138: Sentence is cut-off. Thank you, typo will be fixed.
Ln 184: “Includes” should be lowercase to match formatting of other bullet points. Thank you, typo will be fixed.
Ln 216-217: “In addition it provides information on which surface did the modern carbonate grow and what was its mineralogy” – this feels strangely worded. I recommend: “In addition it provides information on the surface substrate and mineralogy of modern carbonate growth.” Thank you, suggestion will be implemented.
Ln 216-219: “Here modern carbonates refer to…” is repeated twice. Thank you. Repeated sentence will be deleted.
Ln 227: Last sentence (“The colors of the format of the particular field.”) is incomplete. Thank you for the comment, the sentence will now be extended to: The colours refer to the format of that field: Enum, Int, Varchar, Double, or Decimal.
Ln 231: Missing parenthesis after Supplement. Thank you, typo will be fixed.
Ln 363: Remove extra parenthesis before citation. Thank you, typo will be fixed.
Table S1 Column H: “Humidity” misspelled. Thank you, typo will be fixed.
Table S1 Column N: “Dripwater” should be two words here to be consistent with its other appearances in the paper. Thank you, typo will be fixed.
Citation: https://doi.org/10.5194/essd-2026-31-AC2
-
AC2: 'Reply on RC2', Pauline Treble, 31 Aug 2026
Data sets
SISAL_monv1: a global database of cave monitoring observations Pauline C. Treble et al. https://repo.researchdata.hu/privateurl.xhtml?token=43a43257-f06e-4dc5-9e96-6603eabe775f
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 313 | 103 | 23 | 439 | 22 | 25 | 25 |
- HTML: 313
- PDF: 103
- XML: 23
- Total: 439
- Supplement: 22
- BibTeX: 25
- EndNote: 25
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents SISAL_monv1, a global cave-monitoring database. I find this to be a valuable and timely dataset paper. The database will be highly useful for speleothem science, palaeoclimate reconstruction, karst hydrology, groundwater-recharge studies, and data-model comparison. The manuscript is clearly written, the database structure is well documented, and the community effort behind the compilation is substantial. I strongly support publication of this manuscript, subject only to minor clarifications that would improve usability and prevent possible misinterpretation by future users.
1. The database includes site-specific precipitation amount and precipitation isotope sample data, but it also includes monthly precipitation and precipitation-isotope values extracted from gridded products such as MSWEP and IsoGSM. I suggest that the authors explain this relationship more explicitly. In particular, it would be helpful to clarify whether gridded values are extracted at the cave coordinates or precipitation-monitoring-site coordinates, how users should handle differences in temporal resolution between observed samples and monthly products, and whether monthly precipitation isotope values are precipitation-weighted means. This would help users avoid treating local observations and gridded estimates as interchangeable.
2. It would be very useful for users to quickly know the duration of the monitoring data available for each site and each data type. At present, users may need to inspect the sample tables directly to determine the start and end dates of precipitation isotope monitoring, drip-water isotope monitoring, drip-rate monitoring, or modern carbonate sampling. I'm not sure if it is easy to add summary fields, either in the metadata tables or as an additional summary table, giving the monitoring start date, end date, duration, and possibly number of samples for each site and data type. This would greatly improve the discoverability and usability of the dataset, especially for users interested in comparing records across sites or selecting datasets suitable for time-series analysis.
3. Missing parentheses around Lines 363-364.