Maurischat et al present a detailed investigation of dissolved organic matter composition and transformation in various environments on the Tibetan Plateau, focusing on three catchments of Lake Nam Co. The authors find that local characteristics, such as small-scale catchment properties, dominantly shape the DOM composition. It is concluded that there is no single DOM signature, that is generally representative of such a high-alpine region. Instead, many processes along the fluvial-limnic continuum together shape DOM composition.

The study is both interesting and well conducted, the methods and data processing are generally sound. However, the presentation of the findings needs to be improved. In its current form the manuscript text is in parts lengthy, too descriptive and lacking a central theme. As a result, the key findings of this study hardly become evident to the reader. Furthermore, the mixing and transformation processes exemplarily described for a small mass range in Figure 6 need to be scaled to a wider mass range to be able to support the conclusions, which are being drawn.

Detailed comments:

Abstract general comments: The abstract is quite long. In parts, it currently reads more like a results summary. Consider shortening the abstract to allow for a stronger focus on the key messages your manuscript should convey. Some suggestions below

26 – The TP is sensitive to climate change, but this factor was not directly investigated. Could be removed

27 – Carbon cycling at the TP is complex, but expect for the glaciers, all of the examples provided in this sentence occur similarly in many terrestrial ecosystems globally. Could be removed

30 – which network? This word is not introduced and is never used again throughout the manuscript. Could be removed

31 – The information content of the sentences in lines 30-34 is very similar and could be merged.

33 – “headwater-fluvial-limnic pathway” Try to find easier wording to make the abstract more accessible to a diverse audience.

33 – “terrestrial change” this term is not introduced and therefore not clear. Do you mean climate change or land use change or vegetation change?

38 – “endmembers” this term is not introduced yet at this point in the abstract and is therefore unclear. Try to introduce the key terms in the first few sentences and stick to this wording throughout the whole manuscript text. Also, this is the first sentence that actually deals with one of your findings. Not only is this quite late in the abstract to introduce a finding, but the finding presented is also quite weak (not really surprising that DOM differed between endmembers, streams etc). Rather start with your most important finding to emphasize your key result as much as possible.

39-49 This part reads like a results summary. Try to be more concise and less descriptive. Highlight the most important take-home messages.

50 – The first half of this sentence seems like a major finding to me. I suggest moving this part up. At this point in the abstract, too many people might have already stopped reading before they reach the important part.

54 - 56 Again two important findings that should be emphasized more and be rephrased in a way that is easy to understand for a diverse readership.

57 – Your sampling does not target “their susceptibility to changes in water quality and OM cycling”. I suggest rather forming a conclusion that is more directly based on your study design.

59- Graphical abstract looks nice. It could be useful to add a legend specifying which words specify a source, environment, endmember or process, potentially in relation to the statistical groups you test later on.

Introduction

62-71 Most of the issues explained in this paragraph do not strongly relate to your findings. This is a problem, because it leads the reader to expect answers to questions that are not addressed in the sections that follow. The global importance of the TP will be well known to most readers of this journal, so a few sentences on major highlights will suffice. Rather use the opportunity at the beginning of the introduction to bring up the questions/issues you will actually answer later on.

69 – “peak water” this term lacks an explanation and is never used again in the manuscript. Could be removed

70 – “up- and downstream societies” this term lacks an explanation and is never used again in the manuscript. Could be removed

74 – Either you introduce the hypotheses here or at the end of the introduction. Both options are fine, but this split is confusing.

74-76 – Your perception of the marker concept does not become clear here. DOM marker studies commonly use correlations of sum formulae along gradients, like Ideg or Iterr, which you use. But your study does not present such an analysis. I assume you mean that DOM composition reflects differences between ecosystems and indicates the action of certain transformation processes. This is likely true to some extent, but is far from being an actual marker. This should be rephrased or removed.

76-79 Considering your actual findings, these sentences seem like overstatements. While it is true that if one could use DOM as a proper marker for all these environments and processes, this would be very helpful. However, the actually presented results hardly allow conclusions to be drawn on the high level that is stated here. This part should be rephrased or removed.

85-86 “This led to analytical advancements…” This makes it seem like FTICRMS was developed specifically to solve DOM problems. Should be rephrased

91-99 This is an important section, because for the first time in this introduction the highly relevant study design and sampling scheme are discussed. The introduction could benefit a lot from a stronger focus on the aspects discussed in this part, also by moving it more to the beginning of the introduction.

102-118 – Here you pose two broad open questions, which are followed by four lengthy hypotheses in addition to the hypotheses you have already stated earlier. This seems excessive and indicates a lack of focus in presenting your key findings. Try to condense these questions and hypotheses into a few key aspects, not more than 2 or 3. Use short sentences and easy-to-understand language. Make sure they are directly and understandably linked to the topic you introduce in the first 4-5 sentences of the introduction.

111 – “recalcitrant molecular formulae” The concept of chemical recalcitrance is heavily debated in the terrestrial DOM community. This discussion currently not addressed at all. In addition, your interpretation of the word recalcitrance is not explained anywhere in the introduction. (e.g. https://doi.org/10.1111/j.1365-2389.2006.00809.x, https://doi.org/10.1002/jpln.200700049). If you would like to rely on this term, its concept needs to be addressed with clear explanations and relevant literature.

117-118 This hypothesis is so detailed and specific, it reads a bit like it could have been generated after the findings were already known.

Materials and Methods

121-141 Since the whole study is centered around lake Nam Co, it could be reasonable to move the lake description from lines 131-141 to the top.

169 – Figure 1 The green sample category “wetland water” is barely visible on panel a. The map in the bottom left is not discernible and needs to be redrawn with better readability and the tiny words removed. Proper country borders could be useful instead. It also currently does not have a panel label.

Results

274-277 I appreciate you checked for that. Not many studies report on this issue and it’s nice to see you did.

279-367 This section could be shortened to some extent and focus more strongly on key results or aspects that the readers cannot equally gather from just looking at the tables themselves.

344 – Figure 2 – The color palette is not suited for color blind people. Your findings could become more accessible if in addition to the names, their major characteristic was added to the x axis labels. Most readers will not know what the names represent. Just one or two words for each site could make this figure a lot easier to understand. This similarly applies to the text discussing these results. By referring less to the site’s names and more to their characteristics, readers could have an easier time following the results and understanding the conclusions that are being drawn.

367 – Figure 3 – The x-axis title is not really needed here.

369-393 Many of the external variables, that were fitted in this NMDS, were not previously introduced. This section could benefit from its aspects being explained to the reader in the introduction. Currently, many of these parameters seemingly appear out of nowhere. Let the reader know from the start why the factors you test are important and what they represent.

408 – It does not become clear how this conclusion is formed “higher complex compounds originating from plants and soils dominated the DOM of Qugaqie”. In the preceding sentence a low degree of C oxidation is specified, but this is likely not a universal identifier of plant and soil carbon. Please elaborate on this part of the discussion and provide appropriate references to support your conclusions. It seems it is mostly based on the D’Andrilli et al 2019 study, but it not sufficiently explained what they incubated and how their findings help to interpret yours.

413-422 Figure caption: spacing before and after “=” is currently not uniform

427 – word missing? Low molecular-weight/mass compounds?

439 – Please be careful when referring to chemical recalcitrance in terrestrial environments. Various processes can lead to the persistence of some sum formulae (e.g. https://doi.org/10.1038/s41561-019-0417-4). Finding a different wording that explains your observations, but also accounts for the ongoing discussion on persistence/recalcitrance in the terrestrial environment could help circumvent this issue.

446-449 I cannot follow your line of argumentation here. All catchments, even though they are different, show similar amounts of allochthonous DOM. How does this indicate that there is a land use control on DOM composition? I suggest this section to be expanded with more explanation to support your conclusion. In general, the discussion section could become more accessible if more text was spent on guiding the reader from the interpretations of the results to the conclusions. To compensate, some descriptive parts could be shortened.

449 – This is the first and only time the word plagioclimax is used. Needs some explanation or rephrasing

483 – Chemical recalcitrance in soils, same problems as previously mentioned

488-490 This conclusion is currently little supported by the results discussed above. DOM from a spring represents groundwater and is in exchange with DOM from the degraded pasture and ends up in a stream. Along its path, it is highly transformed. This seems reasonable, but does not relate to DOM being a marker. A marker needs to be clearly defined (one or multiple compounds) and show a distinct relationship with the gradient or effect it is supposed to represent. I currently do not see an analysis like this in your manuscript and consequently suggest to abstain from the marker wording.

497- This statement is likely true, but needs a little more explanation. You have studied these sites intensively and likely know all the relevant pieces of information about them by heart. But the reader does not. Try to guide the reader and explain how you come up with your conclusions more thoroughly. Even though you likely mentioned everything already, the manuscript text is long and many readers will not read through the whole text. Try to frame your interpretations in a comprehensive, yet easy-to-understand manner.

498 – “stream samples cluster widely around the NMDS”. There is no cluster analysis being performed. Please find a different wording.

512 – Figure 5 – This figure is really hard to discern. Are all of these highly similar plots needed? If yes, try to highlight the lines you plotted for guidance more strongly. If it is really just about the regression lines, maybe the figure can be summarized somehow?

535 – Exemplary mass-to-charge range: On which basis was this range selected? There is currently no explanation, which makes it seem like a random selection. The small width of this range seems not representative of the DOM composition of the samples as a whole and is correspondingly not suited to inform about mixing processes unless you can argue in detail why especially this region is more meaningful/representative than others. If the mixing process truly appears throughout the whole DOM spectra or at least a large range, it would be nice if this was shown. Something like a Venn diagram could potentially be useful to inform about mixing. Also, this analysis seems more like a result of its own and could therefore be moved to the results section.

545 – “becomes visible in the excerpt mass spectra”. This section needs to be reworked as there is currently no directly visible link between the indices you discuss and what is shown in Figure 6.

564 – The comparison of lake water DOM to marine DOM seems of little use here. The TP is far away from the sea. I suggest referring to ecosystem characteristics only, without implying actual similarity between marine environments and Lake Nam Co.

565-566 – Your sampling scheme can hardly imply millennial scale stability, same problems regarding recalcitrance as discussed above.

566 – Again, the comparison of Lake Nam Co to the marine environment or large arctic rivers seems of little use here. Instead, try to compare with more similar environments such as other lakes. There are many comprehensive DOM lake studies around. The funding statement shows your study is part of TransTiP. There are other biogeochemical lake investigations of the TP and surrounding areas that were conducted in this and related project initiatives. Lake Nam Co is very highly studied using a wide range of biogeochemical analyses. Use this to your advantage. It would be really interesting to compare your findings with them and validate your DOM interpretations this way.

570-571 It is currently not clear how the conclusion was formed that the lake is not influenced by its inflowing streams. This is a strong claim, for which there is currently insufficient support and explanation. The DOM sampled from 30m depth in the lake is highly processed compared to the stream DOM. This is true, but it does not imply that the lake as a whole is not influenced by the inflowing streams.

573-626 The conclusions section is in parts rather a results summary and could be shortened.

609 – smaller lakes were not investigated here, therefore the comparison is difficult

621 – comparison with open ocean signatures seem far fetched

625 – The claim of DOM composition as a proxy that is ready for implementation in monitoring suffers the same problems as the marker discussion before and seems little supported by the current data analysis.

626 – Your investigation highlighted the importance of local effects (line 50). This is in contrast to the conclusion that Nam Co DOM can be used as a proxy for processes occurring throughout the larger southern TP.

The authors present and discuss the quality and composition of dissolved organic matter from the Tibetan Plateau (TP), with a focus on different landscape elements and how they contribute to the divers DOM pool in the terrestrial-fluvial corridor. They build primarily on earlier works from the same group with a similar focus but now add high-resolution MS sourced information on DOM composition. The authors build on a set of around 50 samples to answer how source and the transformations contribute to DOM quality and place these processes in the context of current discussion around land degradation and farming practices in the TP.  The manuscript therefore adds an interesting aspect to an important topic of current relevance. The study is generally sound and well conducted. However, some of the claims and hypotheses are not entirely in line with data, especially given the low sample size in what is referred to as “endmember” systems. It will require more cautious interpretation and description to make these research aspects and findings consistent with the empirical basis that the dataset is offering.

Major points

Much of the manuscript revolves around the contribution of different landscape elements, including glaciers, but also groundwater spring and wetlands. This central idea of the manuscript is perhaps best reflected by the choice of hypothesis 2, that endmember DOM pools contribute unique formulae to the DOM pools in the interconnected corridor. I find that this aspect of the paper has two shortcomings: first, I don’t see how this hypothesis could be falsifiable, and 2. The way this hypothesis is addressed is by comparing single samples from 2 of the 3 “endmembers”. The authors later note that they don’t use factor levels with such limited sample size for comparisons, but then do it regardless, even in situations were it is not entirely clear what statistical significance tests refer to (e.g. L311). I think this requires some carful reassessment of the claims and/or the storyline of the manuscript.

To me, hypothesis 3 and 4 seem repetitive and again focus (at least within this paper) on the comparison of a single lake sample with fluvial DOM. This misrepresents and undervalues the empirical base of the manuscript. The hypothesis of in-stream DOM transformation could, for example, be much better addressed by within-group comparison of the many fluvial samples with respect to in-stream condition, residence time etc, which would much better draw from the strength of the manuscript. In the current state, the reader will expect a different set of samples from reading abstract and introduction and I think this needs some attention.

The authors mention in their last and concluding sentence that “DOM properties have proven here as a selective proxy” for degradation. While that would be an interesting application of HRMS DOM information, I can’t see from the manuscript how such data could be used to draw conclusions about degradation state solely from the fluvial DOM composition, and if this is “the” central takehome of the conclusion, the authors should provide the necessary context elsewhere in the manuscript.

There are further minor points which I hope will be useful in the revision

Method aspects and description: A few aspects seemed unclear in the methods section: L213 “sum of peaks” probably refers to peak intensities, L242, scaling from 0 to 1 and expressed in percentage seems like a redundant practice.

L243 “The three catchments:” is confusing, either remove or place catchments in parentheses

L264 indicate criterion for co-correlation

L279 repetition of information just a few sentences before

L310 decide and harmonize IoS subscript style across manuscript,

L321 In my opinion there is no such relative amount described as “five times fewer”

Table1: it seems to make more sense to include sample number (n) here rather than in the figures

Figure 5: leaving out axis denominations doesn’t seem like a good solution. Consider revision

L534-Figure 6 Selective degradation of DOM is one of the central topics of the paper and can maybe be illustrated in a less raw-data format

Figure 7 Unclear to my why the different “processes” such as pasture degradation is indicated by arrows of different form and direction.

L616 What information are there to estimate residence time in the streams. Could high turbidity not pose a system in which DOM sorption and desorption processes contributes to a specific transformation of the DOM pool in such systems?

L570 “not influenced by inflowing DOM” seems like a far-going claim for an endorheic lake. For example, some of transformed DOM would still be imported in the first place. I suggest to balance this argumentation.

L605 I don’t understand these claims about wetland DOM and don’t see that any of the statements about wetland DOM drivers are supported by the data of this manuscript. Revise accordingly.