The manuscript "Distinguishing vegetation and soil component of d13C variation in speleothem records from degassing prior calcite precipitation effects" by Stoll et al. presents a high quality scientific contribution to the interpretation of speleothem proxy records, in particular carbon isotopes, as well as the trace metal ratios (Mg/Ca) and (Sr/Ca) also explored in the manuscript. It is within scope of Climate of the Past for the development of methods and understanding of speleothem proxy data, as well as some new data in the case studies. The results and interpretation well-argued and presented in an appropriately structured way in terms of text and figures/tables.

The approach is novel and the methods valid. The manuscript is particularly well-presented in its presentation of the complexity of carbon isotopes and PCP, providing a useful review of these proxies as well as a novel method towards improved quantification of proxies. The paper also well-outlines the uncertainties in the approach and then develops a method to numerically incorporate these uncertainties into the quantification of the impact of degassing/PCP on speleothem carbon isotopes. It is shown that while the uncertainties preclude a reconstruct an actual soil CO2 d13C value, it is shown that the method may still be applied to examine the impact on the trend of the reconstructed soil δ13C compared with the measured δ13C values in the speleothem. This is the strength of the paper and is well-illustrated in the case studies. I appreciated the detailed appraisal of the method and output.

As well as providing a method for correcting the PCP impact on d13C via Mg/Ca, my impression of other strengths of this paper are also is in its demonstration that speleothem DMg may vary within a speleothem time series and that this may actually be expected to be common. I think that this point is an important contribution to the community and it is well-explored in the paper, although please see point below about fabrics and flow paths. The other major finding that I think should be useful to the community is that the soil PCO2 and PCP processes can have counteracting influences on the resulting stalagmite δ13C and produce observed nearly constant δ13C timeseries during significant climatic transitions. It is demonstrated in the two case studies and provides good motivation for others to investigate their datasets with this method.

There appears to be cave monitoring paper from the same cave(?) in review (Kost et al)  that this was not cited in the reference list. If it were (and if in open access review?), it would have been interesting to see whether these monitoring data support the approach presented here. Hopefully, this is covered in the Kost paper.

As outlined above, the authors have well-outlined the uncertainties in their approach, although there are two other possible processes that were not or perhaps under-acknowledged. One is the the recent findings by Frisia et al 2022 in Quaternary Science Reviews. This paper showed that Mg/Ca in the calcite may vary with fabric porosity. This will affect the partition coefficient, DMg, which feeds into the equations for correcting PCP impact on d13C. This should be included and the findings of Frisia et al paper also implies that the Stoll et al manuscript would benefit from a description of the fabrics and whether they differ within a stalagmite over sections where DMg is thought to have varied. Similarly, karst flow paths received one brief mention around line 310 yet there has been much discussion in the literature around flow paths and PCP. The manuscript should include a few more sentences around this, rather than focus on PCP has just being a function of CO2 gradients and drip rate. Finally, the fabrics and even speleothem morphology can provide a useful indication of past drip rates and this is covered in multiple other studies and in the Speleothem Science book by Fairchild & Baker. Seeing as the influence of drip rate on PCP appears to be important yet we cannot know how drip rate has varied in the past, some description of the fabrics and speleothem morphology could have been helpful as these can be an indication on whether drip rate has varied in the past.

Other points:

More correct wording of the title may be: “Distinguishing soil component of d13C variation in speleothem records from degassing prior calcite precipitation effects”. The approach is not able to specifically isolate vegetation d13C from the soil d13C CO2 pool and does not discuss vegetation-derived d13C CO2 at any length in the manuscript.

L19: the term “overprinting” is used to describe PCP on speleothem δ13C here and throughout. Suggest “contribution” would be more correct term.

L22: “universally”, this is not a global study so the conclusion can only be that PCP is not the dominant control on this particular study site. Suggest reword.

L28: this sentence is unclear and needs rewording, suggest: “During glaciations, calculated initial δ13C implies trend of increasing respiration rates and higher soil CO2, despite the interpreted reduced drip flux to favour more extensive PCP”?

L32: Why not the lower latitudes? Doesn’t Figure 4a contradict this introductory sentence?

L39: root depth and soil moisture content also contribute.

Table 1: looks like a copy and paste mistake for scenario A1 and A2 as identical definitions given for these variables. This made it a little hard unfortunately to understand some of the distinction in the modelled results for these two scenarios although I don’t think was an important factor in being able to follow the main findings of the paper. But needs fixing here. Also, Delta Ca should be Delta44Ca.

L167: suggest precipitation rate rather than growth rate dependent, i.e., the loss of calcium is not necessarily related to growth rate. Growth rate is usually used to define the linear extension rate in the vertical axis of a stalagmite.

Sub-heading 2.2.3. Are there a few more papers on D44Ca and PCP worth citing here?

L216-217: is this statement supported by the back modelling in the results?

L222: please add the mass spectrometer specifications.

L231: please reword as it is unclear as to whether there were five repeats on each aliquot or whether there were five analyses performed on each stalagmite?

L240: preferred value rather than preferred measurement?

L253: please refer back to equation 3 here.

L300: La Vallina Cave?

L300 paragraph starting here becomes confusing as total range and “fold” range not the same thing. Confusion ensues here for several paragraphs until the reader sees the definition of range used is the maximum/min rather than max minus min value. This definition does not come until the caption in the following table. Please define in the text to assist the reader.

Fig 4 caption: for soil moisture which direction was the experiment? Do the green crosses indicate increasing or decreasing soil moisture?

L432: there should be some earlier description of these dripwater measurements and their method, seeing as the Kost et al paper is not yet published.

Reference:

Silvia Frisia, Andrea Borsato, Adam Hartland, Mohammadali Faraji, Attila Demeny, Russell N. Drysdale, Christopher E. Marjo, Crystallization pathways, fabrics and the capture of climate proxies in speleothems: Examples from the tropics, Quaternary Science Reviews, Volume 297, 2022, https://doi.org/10.1016/j.quascirev.2022.107833.

General comments.

Unraveling paleoclimate signals from speleothem geochemical proxies is an important challenge to address in order to maximize applicability of such proxies. This manuscript combines TE/Ca and calcium isotopes to act as a proxy for degassing and prior calcite precipitation (PCP) and remove the effects of those two processes on stalagmite d13C records. This in turn allows one to interpret stalagmite d13C time series in terms of climate-driven variations in vegetation and soil. The authors use coeval speleothems from a single cave to produce several records to apply their models. A key part of the approach is separating the temporal trends in PCP from those due to soil/epikarst components within a stalagmite. The results show potential for d13Cinitial records to show climate shifts, and that such d13C records are difficult to use without multiple other proxies in the same speleothem. The results for multiple coeval stalagmites indicate that the δ13Cinitial approach may resolve soil/vegetation changes driven by regional climate changes. 

                The manuscript discusses changes in Mg/Ca and Sr/Ca (TE/Ca) in terms of two principal processes - either PCP or changing partition coefficient. The possibility of water-rock interaction (WRI) is offered as an afterthought, without adequate justification for largely discounting it as a key process to be considered. The impact of WRI on TE/Ca, and therefore on unraveling the extent of PCP, is not well addressed. It’s not until page 11, line 305, that WRI is brought up as a possible factor, where it is stated that WRI is not necessary to attain the Mg/Ca range observed for the stalagmites. This misses a key point, however, that WRI – while not necessary – cannot be ruled out as an important process in karst water evolution. In this case, d13C evolution by WRI in the epikarst may occur, and change the interpretation that assumes that all above-cave processes that change water DIC d13C are due to changes in climate/vegetation/soil.  There are examples in the literature of how the multiple effects of WRI and PCP can be parsed out in well understood karst systems. Quantitative approaches to this using TE/Ca and Sr isotopes can be found in Wong et al. (2011, GCA) and Sinclair et al. (2012, Chem. Geol.) There are several places where this issue is further explained in the Detailed Comments below.

The writing and referencing need work, including adding sufficient information and citations to support the points being made. The presentation and language can be clearer to explain the many new model terms used. Similarly, the figures can be improved, such as by adding error bars to age determinations and by adding more multivariate analysis for the scenarios presented. I point out some examples of this in the Detailed comments below, but my comments are by no means complete in this regard. The example suggestions for the writing are aimed at making the scientific message clearer so that readers can get the most out of this study.

There are 32 individual graphics within the figures. The impact of the paper does not justify this volume of material for the reader to go through. I think these and the text can be reduced by 25%.

In summary, the approach presented is a potential useful step forward for the application of speleothem geochemical proxies. A more thorough exposition of alternative processes, such as water-rock interaction, and improved presentation is needed.

Detailed comments by line #.

Line 20. “the initial δ13C of dripwater”. Clarify if this the dripwater once it enters the cave, or once it leaves the soil?

1. Are the 8 stalagmites from 8 different caves? From different climate regimes? Different time periods? Why were these 8 chosen?

23-4. ”Fraction of initial Ca remaining at the deposition of the stalagmite layer”. How is the initial amount of Ca in fluid determined? Is this the same ‘initial’ as line 20 ‘initial d13C of dripwater’? State more clearly.

24-25. State the magnitude of the uncertainty so the reader knows at the outset how well constrained (or not) this value can be using your method.

1. “spanning THE 94 to 82 ka interval"
2. ” higher interglacial dripwater situation”. This phrase does not make sense, rephrase.
3. End abstract with summary of how/where/when other speleothem studies can use this new approach.

33-35. It will be easier for reader to unravel this building block information for the rest of the paper if this is broken up into two sentences.

40-44. This section summarizes the foundational work of Mickler et al. 2004 GCA, 2006 GSA Bulletin, and as such should cite these.

1. Should read 'Superimposed on soil and karst processes are in-cave processes..."

51-3. Sr/Ca dismissed by omission here without explanation.

1. Eq. 1 should be checked for balance, parens, subscripts, etc.
2. Move Sade et al., 2022 citation to end of previous sentence.
3. ‘preserve’, not ‘conserve’

76-8. Be explicit in the description here. Given the quantitative, sequential, and process nature of your model – it would be clearer to explain that the Ca allows estimate of the extent of calcite precipitation, which in turn allows estimate of the degree of degassing.

83-6. More than refer to A, define it explicitly. What is a “degassing slope”? Explain here and in Table 1 definition of ‘A’.

Figure 1b. The y axis label does not indicate what phase the d13C is calculated for.

1. 'landing' better than ‘impingement’, which is not the proper term here.
2. 'partition', instead of ‘partitioning’, change throughout manuscript

119-20. “Such a dependence would serve to amplify the Mg/Ca due to increasing PCP”. This is an important point, but it is stated cryptically. Be clearer.

126-7. “Deviations from expected PCP control may reflect variation in the partitioning coefficients of Sr (or Ba).” Alternatively, such deviations may reflect other processes such as water-rock interaction (Sinclair et al. 2012, Chem. Geol.).

1. need a comma before ‘which’
2. ”DSr increases with increasing calcite Mg/Ca ratio”, see study by Mucci and Morse, 1983, GCA 47, 217.

Throughout manuscript – use convention of notation for partition coefficients with element subscripted

Page 6, first line. See general comments above about controls of mineral-solution reactions (WRI) and host rock compositions as controls on TE/Ca values in drip waters as an important alternative mechanism. “MgCa” should read “Mg/Ca”.

1. “fractionation of Ca”, do you mean fractionation of Ca isotopes?

154-7. How is this known? Refer to a reference.  Add definition of fract factor to Table 1 and refer to that here.

1. “must be assumed…”. Or, estimated from farmed calcite and dripwater from other sites in the same cave.

174-8. To make the case for growth rate control, give the growth rates for the stalagmites from the different settings. Alternatively, if growth rate is not important, then do not provide this info for any of the stalagmites.

1. “good quantitative agreement in estimated fCa_δCa and fCa_MgCa”. What estimates is this based on?
2. “temperature and dripwater initial oversaturation may not have experienced significant variations”. What is this based on?

178-80. How do you rule out variations in water/rock contact times?

1. Abstract states '8' stalagmites not 9
2. Kost et al reference not in reference list, no date given.

215-217. This is an important point, so show the data or refer to the figure of the time series that support this statement. Without this constancy of bedrock dissolution source, the model does not work, correct?.... because it is only accounting for PCP and initial soil variability and not WRI and not host rock TE/Ca variability?

1. Show the age model constraints on the figures as error bars. Indicate such analytical uncertainty on figures or in caption for all figures displaying analytical results.

240-4. This is a run-on sentence and difficult to follow.

1. may also ‘be’ calculated

295, and Figure 3. All but one of these data sets are comprised of only two samples, so it is of limited value to consider correlations here. Is there an overall correlation among sites, which would provide a larger data set?

1. “The 1.7 to 3.5 fold range could be fully explained by PCP”.
   
   - Alternatively, could these values be explained by water-rock interaction? This should be addressed in the text.
   
   - Is it the range of absolute values that is 1.7 – 3.5? What are the units? This is different than stating 1.7 – 3.5 fold range. There are other places in the text where units can be added for clarity.
2. ‘process’ should be plural.

305-6. Indeed, WR interaction is not required, but it also cannot be eliminated, and as such this process is not accounted for by the model, meaning that modeled d13C soil values may also have a component of WRI not accounted for, which would lead to inaccurate interpretations of speleothem d13C as a proxy.

320, Table 2. How are the Age min and max calculated? Why are these given for Mg/Ca and not Sr/Ca?

1. (e.g., warm climates…
2. Fig. 3b or Fig. 4b?

429-31. Difficult sentence to follow. Also, list ranges of values from low to high.

1. “is not well in measuredδ13C GAE”. Do you mean ‘is not well expressed in d13C GAE’?

Fig. 8e. There is a caption but no figure.

1. “Because the extent of degassing and PCP can be strongly conditioned by the drip rate”.. and drip chemistry

566-8. Alternatively, this could reflect water-rock interaction – see Sinclair et al. 2012.

585-6. write more explicitly. This is cryptic.

1. Mg/Ca is “a” widely…

593-4. How about using monitoring of modern dripwaters instead to use Mg/Ca for quantification of PCP?

589-90. “'feature significant components of wide varying solubility and Mg/Ca which give rise to strong incongruency of dissolution.”  This phrase after the ‘and’ doesn't follow. Mg/Ca of what? this is not equivalent to the first term before the 'and'.

637-642. Why three paragraphs, two of which are just one sentence long? It’s all the same topic.

698-699. The role of ecosystem type in the suggested approach is glossed over.