Dear Dr. Birkholz,

Thank you very much for your responses to the reviewers' comments. I consider that you generally responded to their comments well, but one of the referees has raised the concerns below:

Their comment that the discussion will not change their conclusion is also rather dogmatic. The missing information such as bulk delta13C data could have helped their cause. The use of the polygon is to define the sources contributing to the sediment in the mixture. If the mixture values lie outside the polygon, there is a source missing. In this case rather than look for the missing source, the authors have put forward a hypothesis which has the potential to destroy the integrity of the compound specific stable isotope source tracing technique. The most obvious missing source not modelled in their paper is the methanogenic bacteria in the sediment. If the CSIA data from the 1908-1910 sediment depth is used as the bacteria signature, that closes the polygon and all the rest of the sediment samples fall within the new polygon. This is a realistic solution to the author's problem. They are already postulating an effect caused by these bacteria and I agree that that is the most likely solution. I do not believe that any bacteria has sufficient energy to break the ionic bond between the terrigenous FA and the clay soil particle - in many other studies this bond has remained intact over thousands of years.
Conversely, for fatty acids to bind to the soil, they have to be produced insitu and apparently these methanogenic bacteria can do that. If they are producing the long chain FAs and the terrigenous signatures are not being replaced (as they won't be at that depth in the sediment), the new FA signatures will overwrite them and they will be below detection limit. This is not that "activity of bacteria would alter our isotope signals", which implies breaking the ionic bonds of the terrestrial FA, but simply overwhelming the terrestrial FA signatures with the new bacterial signatures. Expressed in this way, the research is providing new information about the effects of hypereutrophic lake sediments.

Therefore, we would appreciate it very much if you could consider the issues mentioned above and then make a revised manuscript. Once we receive the latest version of your manuscript, I will ask one of the reviewers to review it soon.

Again thank you for your nice efforts to improve the manuscript.

Kind regards,

Koji Suzuki
Editor

Dear Dr. Birkholz,

We received a review of your revised manuscript. The referee was very pleased to find your nice improvements, but has still had severe concerns about the correction of Suess effect and a carbon exchange mechanism to explain the highly isotopically depleted long-chain fatty acid signatures in the deep sediments. So please read the following comments from the expert, and make a further revised manuscript with point-by-point responses.

Thank you again for your wonderful efforts to improve the manuscript.

Kind regards,

Koji Suzuki
Associate Editor

---

Comments and over view:
The authors have generally responded well to the referee’s comments and many of the concerns raised have been dealt with, appropriately. There is obviously a missing source in the polygons and that will affect the model output. It is Okay to state that there is an unknown source, but it would be better to complete the search and identify the most likely candidate. As I see it, the most likely candidate is the sediment bacteria that the authors speculated were exchanging the carbon isotopic signatures of the terrestrial FAs.
Because of the fascination with trying to invoke this new source for isotopically depleted FAs in the lake, the authors have drifted away from their stated aim:
“The need to precisely identify sediment sources is especially important in eutrophic systems to enable efficient and targeted restoration measures.” “While the eutrophication history of the Lake Baldegg has extensively been studied (), an in-depth confrontation of the lake evolution with the recent history of the catchment (including land-use and agricultural practices changes) has not yet been performed.” “Our project aimed at filling these gaps. In this paper, the soil isotopic signatures of FAs characterizing the main land-uses of Lake Baldegg catchment are quantified and confronted to the evolution of the CSIA imprint of a 130-yrs long lake sediment sequence.”
That fascination has led to speculation which is unsubstantiated but subsequently treated as fact. There are also errors in the calculation of the Suess effect.
While the paper is important because of the in-depth study of the long-chain FAs in the sediment, the purpose of the paper has become buried in non-essential discussion and needs to be re-written in places to make it more concise and without speculation.
Specifics
The exchange concept has not been adequately dealt with. The authors are still speculating that a carbon exchange is the explanation for the highly isotopically depleted C28:0 fatty acid signature at depth in the sediment while ignoring scientific reasoning and weight of evidence that it is not.
Because it is important to get this correct, I will explain. Look at the statement on Page 10; line 34 onwards:
“Here the example of van Bree et al. (2018) can be consulted, as they found a compound-specific enrichment of mostly C28:0 FA in the water column. Also the accompanying depletion of, in their case, C28:0 FA compared to the terrestrial C28:0 FA signal is hinting into the direction of an aquatic source. A diagenetic transformation of the FAs isotopic signal can also be speculated for the sediments older than 1940. Such an assumption would mean that these sediments would have been affected by carbon exchanges years to decades after their deposition, during the most severe eutrophic phases of the lake history. Why these exchanges would not have affected the younger sediments remains unexplained. And so far, no cases of such a diagenetic transformation have been described.”
Statement: “C28:0 FA compared to the terrestrial C28:0 FA signal is hinting into the direction of an aquatic source”. van Bree et al. (2018) are not just hinting at an aquatic source – they have documented it as an aquatic source and provided a strong “weight-of-evidence” case to support that conclusion.
Statement: “A diagenetic transformation of the FAs isotopic signal can also be speculated “. This speculation is unsubstantiated and incorrect. There are good reasons why “no cases of such a diagenetic transformation have been described” – because it doesn’t happen. A diagenetic exchange of carbon atoms that changed the isotopic signature of the original molecule is impossible. If it could occur, the original C28:0 molecule would be broken into smaller chain-length pieces and no longer contribute to the C28:0 carbon pool. You could, however, mask the original isotopic signature with a more depleted isotopic signature and the observed isotopic signature would appear to have changed if that more depleted isotopic source was not included in the modelling.
Come back to first principles:
1) The FA exudate from plant roots labels the soil by binding on to it. The plant litter does not label the soil because the FAs in the litter are already incorporated into the plant matter and are not available as free acid radicles to bind to soil particles. Water sorting in lakes and bacterial decomposition processes remove the plant litter carbon rapidly but not the FA carbon bound to the soil.
2) The isotopic signature of the FA in the sediment does not “change” going down the sediment core. Because the FA must be produced close to the sediment particle it is bound to, it is the source of the FA that is changing down the core. Hence, we can use CSIA to identify sources.
3) The source of highly depleted d13C isotopic signatures in sediment is associated with methane production and the presence of methane oxidising bacteria (MOB) at the sediment-water interface. Under stratified water column conditions, the MOB can convert methane into CO2 (with the same highly depleted isotopic signature as the methane) at the oxic/anoxic interface in the water column. [Naeher et al. 2014: Tracing the methane cycle with lipid biomarkers in Lake Rotsee (Switzerland). Organic Geochemistry 66: 174–181] for use by algae such as cyanobacteria, which are known to produce C28:0 FAs, and other microbial processes, which produce long chain FAs – these will all have highly depleted isotopic signatures. Look at fig 10, fig 12 and fig 13 in van Bree et al. (2018). These figures show the timing of the production of highly isotopically depleted C26:0 and C28:0 to be after the collapse of an algal bloom in Autumn when their lake was thermally stratified and had probably achieved an anoxic (i.e., no oxygen but has oxygen containing molecules - SO4, NO3) hypolimnion. While van Bree et al. (2018) have linked the highly isotopically depleted long chain FAs to phytoplankton, other researchers have simply attributed the FA production to ‘bacteria” [e.g., Petrišič et al. 2017: Lipid biomarkers and their stable carbon isotopes in oxic and anoxic sediments of Lake Bled (NW Slovenia). Geomicrobiology 34(7):1-12; Steger et al. 2015: Comparative study on bacterial carbon sources in lake sediments: the role of methanotrophy. Aquatic Microbial Ecology 76: 39–47; and many other papers]. Most researchers haven’t looked at the longer chain fatty acids, which is why this paper is important.
4) The organisms with highly depleted long chain FAs produced in the lake during stratification will deposit on the sediment surface and be buried with the next input of sediment and detritus. That will lock the depleted isotopic signature in the sediment as a marker for that time/depth. As the water quality of the lake changes after the removal of the sewage input, the overlying bottom water became anaerobic (i.e., no oxygen or molecules with oxygen -e.g., SO4 and NO3) depriving the MOB of the oxygen source needed to function. Consequently, the MOB probably became a minor contributor to the long-chain FA pool.
Statement: “Why these exchanges would not have affected the younger sediments remains unexplained”. The above is the explanation - the younger sediment was not deposited when the lake was hypertrophic so there would be little if any affect from the MOB process.
Using this scenario, while sewage was entering Lake Baldegg, it is likely that the new source of long chain FAs, locally produced in association with MOB in the lake by the indirect utilization of isotopically highly depleted methane, has masked the stable isotope signatures of the original terrestrial long chain FAs.
It is likely that the isotopic signature from the MOB activity will remain relatively constant in the lake, while the sewage was being discharge into the lake. If this is correct, the MOB depleted isotopic signature would be the missing source end-member. It then follows that the isotopic mixing between allochthonous and autochthonous long chain FAs could result in the variability of the isotopic signatures observed in the sediment. The degree of variability being dependant on the amount of external inputs of long chain FAs added to the lake. As the water quality in the lake improves with remediation treatment, the methanogenic processes become smaller and the terrestrial signatures mask them as they become dominant.
To determine the amount of terrestrial source in the deep sediment, the most depleted isotopic signatures could be used as a proxy for the in-situ production by the MOB, as the missing source in the mixing model.

This explanation is provided to help understanding of a set of very complicated limnological processes working in highly eutrophic lakes.

Recommendations:
1) The section speculating on carbon exchange should be rewritten without the speculation, or simply be deleted. Acceptance of the van Bree et al findings could be a way forward.

2) The polygon test should be repeated with the isotopically depleted deep sediment signature as a proxy for the MOB end-member and the data re-modelled including that end-member.

Other specific points:
Page 3, Sentence starting on line 15 “Despite the …” is confusing and should be rewritten. Suggested rewording “The introduction of an artificial oxygenation system into the lake water column in 1982 (Stadelmann et al., 2002) lead to the disappearance of the varves from 1995. Despite a strong decrease of P concentrations in the lake to below 30 μg.l–1 as the result of lake external and internal measures, the lake has not yet fully recovered from eutrophication (Müller et al, 2014).”
Page 6, line 3: “mostly grasslands since years,” do you mean “grasslands for many years”?
Page 6, I am not sure why the authors persist with discussing the Suess effect. It is a side issue easily dealt with by applying the correction. It is distracting from the main theme of the paper. It appears as though this discussion is trying to justify the speculation of carbon exchange, which is wrong.
The interpretation of the Suess effect, as presented at bottom of page 6 and top of page 7, is ambiguous. The purpose of the Suess correction in the CSIA studies is to normalise all the historical mixture data to the present day isotopic signatures so that present day land use library samples can be used to identify likely land use sources at different depths (i.e., time before present) in the core. The cited paper (Zhao et al. 2001) correctly states the isotopic depletion since pre-industrial age (generally accepted as 1700 AD) has been 2.5‰. This is the same factor identified by Verburg (2007) and can be calculated by the polynomial equation in Verburg (2007). That equation describes an exponential curve with little change between 1700 and 1840 but faster change from then to the present. The Suess correction required for sediment collected in 1940 is -1.6 ‰ and -1.3‰ for 1965. Those corrections need to be added to those sediment data to enable them to be modelled using present day land use samples as the source library.
The statement on page 7, first paragraph “The resulting maximum effect is a depletion of 0.22‰ from 1840 until 2016 for the soils. For the lake sediments the maximum depletion is 0.16‰ between 1840 to 2010. The older the sediments the smaller is the Suess effect on them.” These values are very different to those required to normalise the sediment mixture isotopic signatures to present day equivalents. This indicates a calculation error that needs to be remedied before any further modelling.
The paragraph continues “In 1965 the calculated Suess effect on the lake sediments is only 0.01‰”
By my calculations a correction of -1.3‰ is required. And more “If we assume the calculated effect on the soils to be uniform and take into account that it is smaller than the analytical uncertainty, and further is based on the estimation of parameters like soil incorporation rate and actual effect of the Suess effect on the plants, we think that in our case it can be neglected.”
The Suess effect cannot be neglected. The calculations used by the authors to generate the Suess corrections are clearly wrong and need to be redone and then the modelling needs to be redone with the corrected data.

Recommendation: The calculation errors must be fixed and the section should be re-written to explain clearly why the Suess correction is being used.
A suggestion for this text is: “The Suess effect alters the isotopic signature of the FAs in the sediment from 1840 by around 2.2‰, which is substantially less than the isotopic depletion measured in the deep sediment. The source of that isotopic depletion is unknown but is consistent with methanogenetic processes in the sediment and water column of hyper eutrophic lakes (van Bree et al. 2018).”

Alternatively, the whole section on the Suess effect could be deleted as unnecessary.
Page 11, line 38: remove the reference to carbon exchange (sec.4.2.)
Page 12, line 9: The microbial activity overprints . . . I agree that this is what is happening. A better word than “overprints” is “masks”. Optional change, not essential but has clearer meaning for an English-speaking audience.
Page 12, Sentences starting line 29: “Potentially the algae or other microorganism responsible for the production of the long-chain FAs was [were] not present anymore after 1940. Following the hypothesis with organic material originated from the riparian zone and/or wetlands, the explanation could be that with time, wetlands, evolved after the lake level changes drained themselves or were drained by the farmers until 1940 for the intensification of agricultural practices".
It is correct that the microorganisms responsible for the production of the isotopically depleted long-chain FAs were no longer present after 1940 because the hypolimnion had become anaerobic to a depth of 40 m, removing the supply of oxygen required for the MOB activity in the sediment. The hypothesis that the organic material originated from the riparian zone is unsubstantiated speculation. If it is feasible, it should be measured and the isotopic values used to support the speculation.

Page 13, paragraph starting line 13: This will need to be re-written along with other sections of the paper to ensure the conclusions and recommendations are appropriate. These should also relate back to the original objectives of the study. The title says "Plants or Bacteria?". This manuscript clearly shows that it is both Plants and Bacteria, and that is a good conclusion. Don't change the title.

Dear Dr. Birkholz,

The Referee #2 was very pleased to see your modifications, but this expert has still some concerns on your manuscript. So we would appreciate it very much if you could revise your manuscript following the suggestions from Referee #2 (see below), and send us the revised paper with your point-by-point responses.

Thank you again for your patience and nice efforts to improve the manuscript.

Kind regards,

Koji Suzuki
Associate Editor

---
Referee 2, revised manuscript 2018_288-version 3
Title: “Plants or bacteria? 130 years of mixed imprints in Lake Baldegg sediments (Switzerland), as revealed by compound-specific isotope analysis (CSIA) and biomarker analysis”
Authors: Marlène Lavrieux, Axel Birkholz, Katrin Meusburger, Guido L.B. Wiesenberg, Adrian Gilli, Christian Stamm, Christine Alewell
Comments and over view:
The authors have largely answered the referee’s comments. There is obviously a missing source in the polygons and this has been discussed with suggested sources, both allochthonous and autochthonous being raised as possibilities. It is Okay to state that there is an unknown source.
The only remaining contentious issue is the question of the Suess effect, what it means to this study and how it should be calculated. There is also some confusion in the authors understanding of how the CSIA data is incorporated into the soil and thus how it is corrected.
Referring to section 4.2.3 The necessity of “Suess effect” correction for terrestrial lipids in lake sediments”
Definitions: Terrestrial lipids are allochthonous, i.e., produced outside the lake. Autochthonous lipids are those produced, by whatever mechanism, in the lake water column or lake sediment.
The lipids of concern are the polar fatty acids, which exude from plant roots and bind ionically to the soil particles. They are sufficiently soluble that they can be moved by infiltration rainwater though the surface soils until they bind to the soil particles. Once they attach to the soil particle, they cannot be removed by natural processes, with the exception of bacterial consumption. Under those conditions the concentration of the FA will decrease BUT the isotopic signature of that FA pool remains unchained. As a consequence of this binding, the FAs attached to the soil particles do not have a turnover time, which would imply that they can be exchanged from the soil particles.
The FAs are not waxes. Waxes, including n-alkanes, are non-polar and do not bind to the soil particles. They are generally not soluble in water but form part of the soil humus. When the soil is washed off the land, the soil particles labelled with the FAs are carried to the downstream deposition zone, in this case, Lake Baldegg. During that transport mode, the light humic/organic component is separated from the original bulk soil and will eventually settle on top of the heavier soil particles in the lake. Because that material is organic it will be decomposed rapidly by allochthonous bacteria. In contrast, the FA labelled soil particles will remain isotopically unchanged although the concentration may reduce over time as they are consumed by bacteria.
The application of the Suess correction can only be applied to the allochthonous FAs, because they are in contact with the atmosphere, which is experiencing the Suess effect. The autochthonous FAs may experience some of the atmospheric Suess effects, but they will also experience the biogenically altered CO2 from sediment decomposition processes and plant respiration. This effect has been correctly discussed by the authors. Since these latter processes have not been assessed or documented, the authors have correctly not attempted to model the allochthonous sources in the historical sediments. (There is no reason why they couldn’t model the contemporary sediment in the lake surficial sediments, but they haven’t.)
With these basic facts as a starting point, the description of how the Suess effect affects the CSIA values and how the correction for this effect should be calculated is misleading at best and is misinformation, which should not be in this manuscript or any other.
Of particular concern is the lumping of the FA component bound to the soil particles with the organic carbon from leaf litter and crop debris. When these are mixed in the soil they will have a turnover time associated with the natural carbon cycle, as described by the authors. However, the authors have combined the two distinctly different forms of soil carbon into a supposed soil label, and that is wrong. Only the FAs bound to the soil particles act as labels. The leaf litter could have been blown into the landscape from anywhere, thus contaminating the isotopic signature of the defined land use. Most importantly, because this organic debris is non-polar, it cannot bind to the soil as a label, as explained above.
Yes, the organic debris will breakdown and yes, there are FAs in the organic debris. The difference is that those FAs are already ionically bound to other organic debris and are not able to bind to the soil particles. If bacterial decomposition does release a free FA, it will be a shorter chain length than the FA labels being measured and will not affect the CSIA values of the long chain length FAs being used as labels.
The use of the Feng 1998 equation is not appropriate for the Suess correction. The equation published by Verburg (2007) should be used. The use of the Verburg equation needs to be promoted as the standard method for Suess correction rather than using older equations that do not have the level of sophistication of the Verburg equation. The Verburg equation shows a present day isotopic depletion of the CO2 in the atmosphere of just over 2.5‰ since the beginning of the industrial revolution, which is generally accepted as 1700AD.
The authors are correctly trying to eliminate the Suess effect as the cause of the much larger isotopic depletion of some of the sediment FA tracers in the lake. This can be done with a simple statement just like that: e.g., “The isotopic depletion due to the Suess effect is estimated to be about -2.5‰ since 1700AD (Verburg 2007) and is, therefore, unlikely to have a substantial effect on the autochthonous production of FAs by lacustrine processes. Consequently, the large isotopic depletion observed is most likely from an unknown autochthonous source or associated with draining wetlands etc, etc . . .” Write it how you want. This is all that is needed in this manuscript.
I strongly recommend that the misinformation in section 4.2.3 is removed by deleting everything from page 9, line 23 starting “We thus want to discuss “.to.” speculative” on page 10, line 28, and rewriting the text in a form similar to that suggested above. Figure S4-7 and Table S4-5 may need to be revised to match this textual change, if they are still needed.
The reason for the deletion is that the Suess effect can quickly be eliminated as a causal influence, as suggested above, and is, therefore, not a significant component of the manuscript. As it is written, it is a distraction and the reader would be justified in asking “Why did the authors put this in, it adds nothing to the manuscript and is very confusing.”
General comments:
Throughout the manuscript, the range between two values has been either a dash “-“ or the word “to”. Because some of the numbers have negative values (e.g., -40.0 - -43.3‰), the word “to” should be used consistently throughout the manuscript.
The abbreviation “FA” has been defined as “fatty acid” in the abstract only. This definition must be repeated in the manuscript text at the first occurrence of FA. The abbreviation should then be used consistently throughout the manuscript.
The author’s proof reading of their manuscript is not good. There are nine references which are not cited in the text and two references that have no year. See “references section below”.
Specific points:
Page 6, Line 12: I disagree with the general statement that “The isotopic signature of the samples older than 1940 and from 1964 - 1972 fall out of the source soils mixing polygon, making the use of a mixing model to quantify the contribution of different land-uses to sediment inputs impossible. Certainly, before 1972, source proportion modelling would be inappropriate until the missing source has been found. However, where the more recent sediment CSIA data falls within the mixing polygon, those data can be modelled, and should be for the completeness of this manuscript. I can recommend the use of MixSIAR, without including concentration or any other “priors”.
Page 8, Line 19: Why would you expect higher fatty acid concentrations in lake sediments compared to source soils? Unless the autochthonous fatty acids are produced in the sediment, where they can bind to the sediment particles, they will deposit on the sediment surface where they can be rapidly destroyed by bacterial decomposition processes. This means that the concentrations in the lake will decrease. Recommend that this sentence and the next are deleted as unnecessary and speculative.
Page 8, line 24: Change “-70+-15‰” to “-70 ±15‰”
Page 8, lines 29 and 30 (and elsewhere): Where did the “n-fatty acids” expression come from. Please remove the “n-“ and just use “FAs” consistently throughout the manuscript.
Page 13, Line 11: The sentence “As expanded above (Sect. 4.2.), a high discrepancy in isotopic values between long-chain FAs of close chain-length points to a degradation of the isotopic signal.”
No it doesn’t point to degradation. It probably points to a variable amount of the unknown source or other sources in the FA mixture in the sediment. The isotopic signal will only change through fractionation, and fractionation will destroy the original long chain FAs producing shorter chain FAs, which are not measured or included in the data. Since part of section 4.2. should have been deleted, this sentence and the following sentences need to be checked against the rewrite of section 4.2.3 for consistency. As written, this appears to be an example of where earlier speculation and misinformation has been treated as fact, thereby perpetuating misinterpretation of the otherwise good data.
Page 9, Line 23: (Verburgh 2007) incorrect spelling of ‘Verburg’
Page 13, line 29: The sentence “The CSIA signals of arable lands as well as orchards plot halfway between grasslands and forests, which may render difficult to correctly attribute the sources of sediment samples lying between grasslands and forests end-members.”
They don’t always plot in the same positions with the different isotopic tracers, therefore discrimination should be possible. Have you tried to model the data using a stable isotopic mixing model, such as “MixSIAR”, just using the contemporary surface lake sediment as the lake endmember? The surficial sediment will have received negligible effect from any possible autochthonous source. Just looking at the isotopic signatures gives an indication of which sources are present. Modelling the isotopic signatures gives a robust assessment of the isotopic proportions. Using an isotope-to-soil proportion converter, as presented in Gibbs 2008, gives the proportional contribution of each source soil to the sediment mixture. These calculations will be valid to a date/depth of about 1970 and would add greatly to the usefulness of this manuscript.
Page 14, line 3: The text “(2) would hint into the direction of an additional source with low C28:0 FA concentrations . . .”. The interpretation of the data doesn’t “hint in the direction of” it ‘indicates an additional source’. That source doesn’t have “low C28:0 FA concentrations” it has ‘a depleted C28:0 FA isotopic signature’. Since concentration is not a factor when dealing with isotopic signatures, the relationship between C28:0 and C24:0 and C26:0 FAs should be checked in the text statements and discussion.
Page 14, line 8: “(iii) algae with depleted δ13C values due to described effects of hydoxilation reaction of CO2 combined with high pH values in the epilimnion and CO2 undersaturation, . .”
This paragraph does not fit what you are describing and should be deleted.
Explanation: The high pH by itself does not alter the isotopic signature of C. High pH in the water column is most likely due to consumption of dissolved CO2 by algae during photosynthesis, which is what you are alluding to. The algal species that cause the highest pH are cyanobacteria, which are bicarbonate adapted so that they can utilise bicarbonate (HCO3) when all the CO2 has been used. Thus, they outcompete non-bicarbonate adapted algae such as greens and diatoms. However, cyanobacteria are buoyant and tend to stay near the surface for high light and, since the majority of the CO2 and HCO3 at the lake surface water will be of atmospheric origin, their isotopic signatures will mostly reflect the atmospheric value -12‰ for fractionation. The cyanobacteria can come into contact with isotopically depleted CO2 when the lake mixes in autumn releasing the nutrients (especially P) and the methanogenically-derived isotopically depleted CO2 that have accumulated in the bottom waters. (Aquatic macrophytes can also raise the pH if they are bicarbonate adapted).
Page 14, line 14: The text “CSIA was proven to be not suitable to quantitatively unmix terrestrial sources from the Lake Baldegg historic sediments, and thus to apportion the relative contribution of different land-uses to the sedimentary archive as long as the isotopic signal of the missing source is not known.”
This is not strictly correct. It was not ‘proven’ that CSIA was not suitable to quantitatively unmix terrestrial sources from Lake Baldegg historic sediment, the authors have stated that, because the isotopic data from the historic sediments did not fall inside the mixing polygon, they would not model them. That is a valid approach but does not constitute proof that CSIA was not suitable to quantitatively unmix, it means that the data was unsuitable to be modelled using CSIA to quantitatively unmix. Very important difference.
In this respect, there is no evidence in the manuscript that the authors have attempted to model any isotopic data, including the data that did fall inside the polygon. This is a missing element that could improve the paper.
Notwithstanding this, the conclusion that the authors see the imprints of plants AND bacteria in the Lake Baldegg sediments is an important finding and that conclusion can be drawn from the data presented without modelling.
References:
Jansen, B., van Loon, E.E., Hooghiemstra, H. and Verstraten, J.M., Improved reconstruction of palaeo-environments through unravelling of preserved vegetation biomarker patterns, Palaeogeogr. Palaeocl., 285, 119-130.
Reference has no date and is not cited in the text.
Jansen, B., de Boer, E.J., Cleef, A.M., Hooghiemstra, H., Moscol-Olivera, M., Tonneijck, F.H. and Verstraten, J.M., Reconstruction of late Holocene forest dynamics in northern Ecuador from biomarkers and pollen in soil cores, 25 Palaeogeogr. Palaeocl., 386, 607-619, 2013.
Reference not cited in text.
Keeling, C.D., The Suess effect: 13Carbon–14Carbon interrelations, Environ. Int., 2, 229–300, 1979
Reference not cited in text.
Krull, E.S., Skjemstad, J.O., Burrows, W.H., Bray, S.G., Wynn, J.G., Bol, R., Spouncer, L., and Harms, B., Recent vegetation changes in central Queensland, Australia: Evidence from d13C and 14C analyses of soil organic matter, Geoderma, 126, 241-259, 2005.
Reference not cited in text.
Réveillé, V., Mansuy, L., Jardé, E. and Garnier-Sillam, E.: Characterisation of sewage sludge-derived organic matter: lipids and humic acids, Org. Geochem., 34, 615-627.
Reference has no date.
Schelske, C.L., and Hodell, D.A., Using carbon isotopes of bulk sedimentary organic matter to reconstruct the history of nutrient loading and eutrophication in Lake Erie, Limnol. Oceanogr., 40, 918-929, 1995.
Reference not cited in text.
Steger, K., Premke, K., Gudasz, C., Boschker, H & Tranvik, L.: Comparative study on bacterial carbon sources in lake sediments: The role of methanotrophy. Aquat. Microb. Ecol., 76, 39-47, 2015
Reference not cited in text.
Suess, H.E., Radiocarbon concentration in modern wood, Science, 122, 415–417, 1955
Reference not cited in text.
Upadhayay, H.R., Bodé, S., Griepentrog, M., Huygens, D., Bajracharya, R.M., Blake, W.H., Dercon, G., Mabit, L., Gibbs, M., Semmens, B.X., Stock, B.C., Cornelis, W. and Boeckx, P.: Methodological perspectives on the application of compound specific stable isotope fingerprinting for sediment source apportionment. Journal of Soils and Sediments, 17, 1537–1553, 2017.
Reference not cited in text.
van Bergen, P.F., Bull, I.D., Poulton, P.R. and Evershed, R.P.: Organic geochemical studies of soils from the Rothamsted classical experiments. I - Total lipid extracts, solvent insoluble residues and humic acids from Broadbalk wilderness, Org. Geochem., 26, 117-135, 1997.
Reference not cited in text.

Dear Dr. Birkholz,

It is a great pleasure now to accept your manuscript for publication in Biogeosciences. Thank you very much for responding to the reviewers' comments so carefully and completely.

Before it is published, please consider my editorial comments below.

P1, L28 and hereafter: please add a comma immediately after e.g. (e.g., ...).

P1, L29: "Sphagnum" should be italic. Also "spec." should be replaced by "species".

P2, L7 and hereafter: please add a comma immediately after "i.e.".

P3, L21: Please add a period immediately after al (Müller et al., 2014).

P3, L27: Please add a space between 2 and m.

P19, L5: The "2" in "CO2" should be subscript.

P19, L5: The "13" should be superscript.

Thank you again for choosing the EGU-journal Biogeosciences.

Sincerely,

Koji Suzuki
Associate Editor