How catchment characteristics influence hydrological pathways and travel times in a boreal landscape

Jutebring Sterte, Elin; Lidman, Fredrik; Lindborg, Emma; Sjöberg, Ylva; Laudon, Hjalmar

doi:https://doi.org/10.5194/hess-25-2133-2021

Articles | Volume 25, issue 4

https://doi.org/10.5194/hess-25-2133-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

Linking landscape organisation and hydrological functioning:...

https://doi.org/10.5194/hess-25-2133-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 25, issue 4

Research article

|

20 Apr 2021

Research article |

| 20 Apr 2021

How catchment characteristics influence hydrological pathways and travel times in a boreal landscape

Elin Jutebring Sterte, Fredrik Lidman, Emma Lindborg, Ylva Sjöberg, and Hjalmar Laudon

Download

Final revised paper (published on 20 Apr 2021)
Preprint (discussion started on 01 Apr 2020)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Jutebring Sterte et al.', Anonymous Referee #1, 06 May 2020
- AC1: 'Response letter to reviewer #1', Elin Jutebring Sterte, 09 Jul 2020
RC2: 'Linking Groundwater Travel Times Review', Anonymous Referee #2, 28 May 2020
- AC2: 'Response letter to reviewer #2', Elin Jutebring Sterte, 09 Jul 2020

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by editor and referees) (09 Jul 2020) by Conrad Jackisch

Dear Elin and co-workers,

Thank you again for submitting your manuscript to our special issue. Thank you also for your replies to the reviewers' comments. You have drafted ways to take up their constructive criticism in the revisions.

I think it is also worthwhile to revisit our abstract of the special issue for that. We hope to find contributions which really aim at the intertwined relations of landscape properties and landscape functioning. As you have discussed in your replies, working out the novelties towards indicators of changes, methodological implications or restrictions etc. would be a way forward.

So far, I feel reluctant to see a shifted focus on baseflow alone to be sufficient. But in combination with your suggestion to work out the role of snow coverage and the interrelation with the catchment's characteristics this appears more plausible to me.

You have stated that the Kryklan catchment is one of the best instrumented and understood boreal catchments. From the point of view of our special issue, this could be transformed into a large advantage to really track the value of this data to improve our understanding. Given the situation that MikeSHE and particle tracking is not the novelty and that it is neither the findings about MTTs/TTDs, maybe turning the argument towards what is the added information and where does it actually come from could be a way to also include a critical view regarding the state of our hypotheses and tools in hydrological sciences. During the discussions among the Guest Editors we found that much of our intentions with the special issue can be seen as closely related to the Critical Zone/Hydropedology concept of Henry Lin et al. but for the landscape/catchment in our case.

I hope you can align your revisions to these thoughts and would be pleased to receive your revised manuscript in due time.
All the best.
Conrad

Hide

AR by Elin Jutebring Sterte on behalf of the Authors (17 Sep 2020) Author's response

ED: Referee Nomination & Report Request started (03 Oct 2020) by Conrad Jackisch

RR by Anonymous Referee #1 (03 Nov 2020)

Suggestions for revision or reasons for rejection

General comments:
The paper has definitely seen some improvements during the last revision. The authors did a thorough job of considering and answering the reviewers questions and concerns. In particular, the introduction is more complete and the methods section contains more important details.
The addition of investigating the seasonal differences has definitely made the paper more interesting and impactful. Stressing the specific local settings (with a long, snowy, low-recharge winter period, mire abundance and regional soil types) and directing the investigation and discussion towards these factors moves the manuscript in the direction of presenting novel results and insights.
The presentation style and structure has, however, suffered from incorporating the new research. Often the manuscript is not that clearly written and the individual sentences are not connected in order to show relationships and tell a coherent story. Regularly there is a lack of precision in the wording that makes it hard to follow the discussion and arrive at an otherwise obvious conclusion.
The authors decided to move away from investigating the entire travel time distributions and concentrate on mean travel times and young water fractions alone. I find this to be a missed opportunity since the use of the distributed model allows for a more in-depth analysis of the catchment system dynamics (and travel time distributions contain a lot more information on that).
The discussion section is mainly filled with an evaluation of the catchment model (2.5 pages) while the new results concerning travel times only occupy 1 page. I would like to see a significant improvement of the way the discussion and conclusion is presented in terms of structure and writing style.

Specific comments:
Key points
Line 18: ‘…while THE FRACTION of mires…’

Abstract
Line 29: Why do you write ‘the following snowmelt’? Following what exactly?
Line 30: Better describe what potentially causes the seasonality in MTTs (changes in precipitation, temperature, vegetation?). ‘Seasonality’ causing seasonal differences is not a good explanation…
Line 32: ‘…(YOUNGER than three months)...’
Line 34: This is the only time in the entire manuscript that you mention the word ‘synchronicity’. This should be explained in the results, discussion and/or conclusion section.
Line 36: What about the mires affected the young water fraction? Was it the fraction of mires?
Line 37: So catchments with mires cannot also be forest-dominated?

Introduction
The introduction has improved a lot. Previous work is now covered more completely and the authors have obviously read and inserted the suggested additional references appropriately. Also the focus and objective of the study has been shifted towards exploring TTDs and their controls in different seasons which adds significantly more relevance to the results.
Line 105-106: A strange sentence: ‘…water that have spent…’, ‘…sub-subsurface…’.
Line 152: ‘…has been included…’?
Line 183: ‘…gave a more precise location…’? What do you mean by this? Better spatial resolution in general? A more precise location of certain features? And what happened to ‘the threshold level of the lake outlet of C5’?
Line 193: What happens to particles that leave the saturated zone but re-enter it later (re-infiltration)? Is that somehow accounted for? If not, in which way would that influence your results?
Line 228: I would prefer a more unambiguous description of the isotopic value. Instead of ‘more depleted’ I would write ‘more negative’ because the signal could be more depleted in the heavy but also more depleted in the lighter isotope.
Line 230: In the equation you write ‘spring/summer’, above and below you only use ‘spring’.
Line 273-274: Does this include snowmelt isotope measurements? If not, a huge amount of water input to the groundwater (with very different isotopic composition) would be neglected.
Line 290: What about the time it takes to reach the water table? What deviations from the ‘real’, ‘complete’ MTT do you expect when using this simplification?
Line 297: Can a standard deviation be ‘high’. I’m not sure – but I’m not a native speaker.
Line 298: The median would also be an appropriate measure that has been used frequently.
Line 301: I am confused because I thought you wanted to look at MTTs for different seasons. What do you mean by ‘…when the median was stabilized for all sub-catchments’? Do you mean that the seasonal change in the median TT was stable for all sub-catchments?
Line 306: Don’t forget the time that water needs to percolate to the water table. This could be significant in certain situations (especially when it’s dry).
Line 338-340: A bit confusing. So now it reads as if first you determined the mean travel time of all catchments then you determined the groundwater mean travel time and then you determined the groundwater geometric mean travel time? It would be easier to understand if you connected the sentences better…
Line 350: Enhance or dilute? These are not antonyms. What do you mean exactly? Also it is strange to mention that they are affecting stream solutes in ‘different ways’ without giving any more details…
Line 353: You are using age and gMTT interchangeably. Please stick to one term for clarity. Also, I think you want to say that the ‘discharging’ groundwater is youngest. There could be very old groundwater in the system that does not discharge to the stream during that period of time.
Line 354: Try to be clearer in your use of language. It is not the groundwater that ‘increases’. It is either the age of the groundwater or the gMTT of the groundwater that increases.
Line 355: Again, more clarity is required. Travel times are not ‘old’, travel times are ‘short’ or ‘long’. Discharging water can be ‘old’ or ‘young’.
Line 361: Why do you write ‘despite a larger young water fraction’? Since this larger fraction is not considered in the particle tracking, it should not have a big influence on the groundwater gMTT. Also, why do you write ‘even more’? The decrease in groundwater gMTT for C4 is smaller than for C2.
Line 390: ‘strong statistical significance’ or ‘significant statistical correlation’?

Discussion
I would wish for a careful revision of the style and structure used throughout the discussion. Often there is a lack of cohesion between the individual statements and it is not explained why statement a) would support statement b). One example of this can be found at lines 471-473.
Line 430: ‘…for the MEAN water age…’.
Line 437: I would add the acronym of the model for better recognition (STARR or something like that).
Line 450-459: I understand your reasoning. However, especially in catchments where the soils have low conductivities, the difference between groundwater TTs and overall TTs could be quite large. Can you at least try to give some estimations on how long the percolation through the different soils to the groundwater table would take?
Line 472: How does the negative correlation suggest that the model produces credible patterns?
Line 484: ‘less’ is not the correct descriptor for a difference.
Line 489: ‘biogeochemical consequences’?
Line 491: In general, try to be more precise. Instead of writing: ‘Inter-annual precipitation and evapotranspiration variability likely caused the relationship to be less evident…’, better write ‘A larger inter-annual precipitation and evapotranspiration variability likely caused the relationship to be less evident…’. This also applies to the variable names – one time you write ‘the summer ΔδO18’, another time you write ‘ΔδO18summer’. Be more consistent.
Line 525: Again, more precision in how you describe your results is necessary: It is ‘the fraction of low conductive soils’ and not just ‘low conductive soils’.
Line 550: They don’t ‘receive’, they ‘produce’. Also, what exactly is the ‘annual age of water’?
Line 555: A ‘similar annual age’? Please be more precise.
Line 558: Does it displace the older water or is it just added to the older water (and mixed)? One would cause the stream water to first become very old and then very young, the other one would cause the stream water to become somewhat younger on average.
Line 559: What you describe is not called ‘transmissivity feedback’. Transmissivity feedback describes the process that discharge from hillslopes increases over-proportionally when the storage in the hillslope increases - due to the fact that soil layers closer to the surface have higher hydraulic conductivity. Sloppy and makes me question the general understanding of the processes taking place.

Figures:
Figure 1: Catchment topography is shown in panel (d) – not (e).
Figure 2: (a) The use of the symbol ∝ is incorrect here. It means ‘proportional to’. The difference in the isotopic signature can only be proportional to the fraction of young water. Also I think you are missing a Δ before baseflow MTT.
(b) You haven’t explained yet what gMTT is. Also, there should not be a line for the precipitation BC concentration (it is not connected to any MTT), you can instead mark it on the y-axis.
Figure 3: Better write ‘(a) Steps of the particle tracking’. In panel (b) write ‘Average depth to the groundwater table’. I don’t understand what you want to say here: ‘Note that the top vertical layering of the saturated zone was set to 2.5 m below the ground surface, and the thickness thereafter follows the soil layers (thickness increasing with depth).’ Can you add another panel to the figure to show what you mean?
Figure 4: The figure shows the ‘age’ of the particles reaching the outlet, not the ‘timing’ as you write in the caption.

Tables
Table 1: What are the numbers in parentheses in the ‘Sorted sediments’ column?
Table 2: One row just reads: ‘Soil frost’. What does that mean?
Table 5: Is aMTT the arithmetic mean of the ‘particle tracking (groundwater gMTT)’ or of the ‘particle tracking adjusted for overland flow according to Eq. (4)’? Please clarify.
Table 6: I guess something went wrong in the formatting of the caption ‘according to a p>0.05, b a p<0.05, and c p<0.01’.
An idea: Since the upper two rows of each table are identical, why not make one bigger table out of the four smaller tables? Something similar to the attached table (or transposed).

Appendix
Line 835-837: What is ‘ai’?
Line 841: ‘is Krycklan’?

Supplements
The supplement was supposed to be deleted, I guess. It is still available with the manuscript though. Don’t forget to delete it, otherwise readers will be confused.

Referee Report: PDF

Hide

RR by Anonymous Referee #2 (09 Nov 2020)

Suggestions for revision or reasons for rejection

I would like to commend the authors on the extensive revisions they undertook in response to the two reviews in the first round. The most significant concern raised by myself, and the other reviewer, was that the organization around the catchment size hypothesis lacked novelty. I commend the authors for taking this comment to heart and significantly revising the text to re-focus their analysis on the aspects of their study that are more novel: (1) their unique study location and (2) the ability to explore seasonal differences in MTTs within this setting. I think this change along with the expanded literature discussion greatly improved the manuscript. And I agree with the authors that this new focus brings much more novelty to the study.

I also appreciate the addition of Figure3. This makes the approach much more clear and addresses some of my concerns about the confusion of the methodology.

My only remaining concern is in the description of the MIKE-SHE model itself. While I realize that most of this is published in other locations, I was hoping that the authors would provide a better description of how the model is actually working -- ie. how physical processes and interactions are represented. Section 2.2 was modified but is still mostly about the parameterization of the layers. I think this section should be updated to provide a more comprehensive overview of the model for those who are not familiar.

Also a more detailed comment on this section: This description of the upper most soil layer thickness (line 175) is confusing and doesn’t match with the table where it looks like this is only true in the case of till? Sandy sediments appear to have a thickness of 0.8 for the top layer? Can you clarify this?

Hide

ED: Reconsider after major revisions (further review by editor and referees) (09 Nov 2020) by Conrad Jackisch

Dear Elin and co-workers,

Thank you again for the strong improvement of your manuscript. I have received and studied the reports of two referees. Moreover, I quickly went through your manuscript myself. Referee 1 raised a series of concerns to sharpen the structure and to strengthen your conclusions with respect to insights about the system – not the model. Referee 2 is more positive about the reached state. He/she points to the lack of structural detail in the model setup description.

I strongly agree with both referees and would like to ask you for another round of major revisions along the suggested lines. In my view, your study has a strong data-related focus, which is somehow difficult to trace and obscured when referring to your MikeSHE model and the added corrections. Moreover, I find it particularly difficult to draw the lines to catchment functioning beyond small vs. large or GW dominated vs. OF dominated. I have full confidence that you can and will turn your manuscript into a nice paper when adhering to and working clearly towards the key messages about landscape functioning.

I hope you find the reviews (incl. mine) helpful for this. If you have further questions, please contact me.
All the best. Many cheers.
Conrad

Here is my review:
L115-124: After studying your methods and consulting your 2018 paper, I am not quite sure how you used MikeSHE and if your application can really hold as physical 3D GW, 1D vadose zone and surface water model. If I am not mistaken, MikeSHE allows to use different approaches to calculate the hydrological fluxes through the different model units – which are not necessarily “physically” defined. Moreover you superimpose mixing assumptions (eq. 3 and 4), which tries to compensate for some model limitations. L123f is full of further claims, which I find difficult to recover in your manuscript. I suggest to really stick to the approach using data from 14 sub-basins and calculating seasonal gMTT from this data and the model and evaluation of this gMTT findings.

Sec. 2.2: Despite being in favour for the reached brevity, the essential points are difficult to discern. Even after consulting your 2018 paper, I cannot really see, if the numerical discretisation and setup is as presented in Bosson et al. (2008) table 3-1 with a more fine vertical discretisation near the surface to enable a calculation of unsaturated flow (see Vogel and Ippisch, 2008, 10.2136/vzj2006.0182). L174 is rather confusing - especially when consulting Tab 2 with the greatest depth of surface soils to 7 m (out of 100 m) and the rare occasions of GW tables deeper than 10 m (Fig 3b).
Also the actually used “physical” concepts remain unclear to me. In L180 you refer to depth-dependent drainage functions with given hydraulic conductivity values (Tab 2), which I could not fully understand from a quick scan of your 2018 paper and Bosson et al. (2008). I am under the impression, that this is some sort of Darcy flux with unit gradient.? Is this correct? How is infiltration calculated?
Please do not get me wrong: I have no problem with your model selection and the used approaches. Moreover I like the brevity. However, I would suggest to clarify which “numerical engine” (to use the term from Graham and Butts (2005)), which concepts and which discretisation is used. Maybe tab 2 could also come as a figure for easier understanding and to get the model out of the way for the more interesting insights?
I got confused by the brief announcement of the particle tracking approach in Sec 2.2 (L190ff.) but the actual subsection 2.4 with the more detailed description. Also giving tab 3 without explanation how porosity affects the particle tracking added to my confusion. I suggest to tackle all particle tracking in sec. 2.4.

Fig 3 is insightful, however I find it difficult to understand the low number of particles. If you get about 20 particles per year and most particles are released rather quickly, can your really robustly sample the tails? I would also welcome a word about the idea behind only “releasing” particles in the saturated zone (and not simply adding markers to all incoming water).

Sec. 2.3: I am slightly confused by the title and the following description of the system’s seasonality. Maybe I went through the manuscript too quickly. I suggest to include a statement about the references (δ18O and BC) and the respective concepts (Fig 2) incl. the site seasonality before going into detail. L125-135 does not really serve this.
With respect to the employed “correction terms” (eq. 1,3,4) I cannot really see how much this influences your results. At the moment it simply resides in the realm of post processing of the MikeSHE model output and the data. However, it comes with a series of assumptions.
For linear regressions (like eq 1) it might be sufficient to name the scaling factor and intercept and focus on what is behind the regression? (Is there seasonality in the signal… Shouldn’t there be a weather dependency of this relationship (depending on actual evaporation)?)
Eq 3 in my eyes says that the flux in a cell is averaged over all surface fractions but that the concentrations of non-mire surface fractions are diluted by the same flux of non-loaded water from mires. I see its practical meaning but I am not quite sure about its theoretical underpinning and if this has effect on the results.
Eq 4 is a similar scaling approach. This appears to have strong implications. After thinking about it I am not really sure what the propose of MikeSHE remains if groundwater flow is indeed calculated as "drainage function" based on conductivity and when overland flow is actually imposed as surface fraction. I assume that your

Eq. 1-4 in general: I am not really convinced by the chosen notation with a mix of variables and names and everything. First of all, the equations are rather simple. Hence I could imagine that they are actually not needed.
The calculation and averaging of Δδ18O is also quite straight forward (eq 2). Generally speaking, the average of an average is equal to the average of the whole set as long no weights are added. Hence I am not sure if this can be presented more easily?

Sec 2.4: I find it very difficult to understand why particles are only traced in the GW and not form the entry into the system. I suspect this to be a limitation of the model. However, this forces the correction with eq. 4.? The second post processing step comes from the geometric vs. arithmetic MTT calculation. Obviously, this has dramatic impact. Using the median or a log(time) conversion could be other approaches. In the later course you refer to percentiles below 1 year… First of all, I would expect a little more reference to the literature about these steps. I suggest to clarify the used methods also with respect to your forthcoming comparison of Δδ18O and BC vs. gMTT. Maybe a more conceptual description about the theoretical similarities about the respective approaches would be helpful to prepare and clarify your findings? I would also appreciate to be able to discern what of your results can be attributed to the model or the decisions in your post processing. As a minor point I was wondering if 1 particle per 20 mm GW recharge is really sufficient? That means that one year is represented by ±18 bins only…

Results: I will go through the remainder rather quickly.
Fig 4: Where is the gMTT, MTT, median, percentiles etc.? How does this compare to estimates from data (before turning to the much more aggregated Fig. 6 and 7)? I think a brief summary of the nice water dynamics performance from your 2018 paper would be worth mentioning. (This could also come as staring point in your methods.)
Tab 5 and Fig 5: I find it not easy to look through these many results with the few guidance offered. Likewise I found it difficult to reconnect your conclusions to the presented graphs and numbers. (L344 how can I see deep and long flow paths?)
Fig 5: What is “discharge fraction” in the main panels referring to? (percentage of annual discharge?) Why is the overall shape of the curves of each GW fraction very much similar although the flow paths should be very different? Would it be an option to avoid the C# but to refer to real descriptors (like whole catchment, small forest subbasin)?
Fig 6: Why is not always gMTT used on x-axis? Would it be an option to give focus to some sub-basins, which I can find in each plot? Eg. C20 is in b, d, f but not in a, c, e. Also the others from Fig 5 are not to be found in all plots. Maybe this could also strengthen your story, when you can adhere to some examples motivated throughout the manuscript?

Discussion: Again just quickly.
I follow referee 1 that the focus should really be the data and the characteristics. When it come to the model, I do not see that there is sufficient basis about the employed concepts and numerics to attribute found behaviour to model specifics. Which are the most valuable and/or most delicate properties which will be affected by changes in snow melt and evaporation? What can we learn for other catchments with far less data than Krycklan can offer?
You state that soil hydraulic conductivity is essential. However, this appears to be somehow one of the few site parameters. I am inclined to claim this importance a model artefact - especially given the rather broad definition (in orders of magnitude).
You tackle the role of mires throughout the manuscript but it remains rather implicit. Maybe this could be worked out more clearly based on a hypothesis? Or is it really a side topic?

Conclusions
You state “Yet, the combination of stable water isotopes, stream water chemistry, and particle tracking provided a consistent picture of how the boreal landscape functions hydrologically and what processes and factors are of importance.“ I am not sure if these threads have been really brought together yet.
Moreover, the role of snow packs is not really worked out in your study (or I have missed it). Since it is also in the title I suggest to really give it more attention.

Minor things I spotted:
L96f. Verb missing?
Tab1 caption: Please clarify "Soil proportion based on the soil"

Hide

AR by Elin Jutebring Sterte on behalf of the Authors (18 Dec 2020) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (05 Jan 2021) by Conrad Jackisch

RR by Anonymous Referee #1 (01 Feb 2021)

Suggestions for revision or reasons for rejection

General comments:
The authors did a good job in improving the readability and overall quality of the manuscript in the last iteration. The authors’ response was considerate and explained the reasoning behind all relevant changes and decisions. The fact that I still have a lot of questions and comments is not a bad sign because I see that my former comments were appreciated and helped in improving the paper. Therefore I am sure that the authors will also be able to use the new review to the benefit of the manuscript.
My main questions that need attention concern the methods section. Some important details are still unclear to me:
• For constructing the TTDs you show in Figure 5 do you take all the particles arriving in a stream at the catchment outlet over the entire modeling period?
• So these are not individual backward TTDs for a single moment in time?
• They are, however, similar to flow weighted averages of all individual backward TTDs, is that right?
• Accordingly, your annual and seasonal MTTs are flow-weighted, right? You take all the particles arriving in a stream within a certain period of time (e.g. spring) and take the average of the distribution.
• How exactly do you account for the reduction of travel times due to overland flow? Please show the equation.
Please make sure to add all this relevant information to your methods section.

Specific comments:
Title
It includes everything. Maybe a focus on the most important issue would be better. Is isotopic composition not part of the stream chemistry? For example, a more logical title would be:
How catchment characteristics influence flow pathways, travel times and stream chemistry in a boreal landscape

Key points
The last key point is confusing. The greatest seasonality is observed in silty soils because they contribute a larger amount of old water during winter baseflow compared to other soils? And why do you mention the mires in the same key point? Are they also related to the seasonality in catchments with silty soils?

Abstract
Line 25: ‘…the pathways and associated travel times of water in 14 partly nested…’. First you use the model to investigate the pathways. Each of the pathways has an associated travel time. This is the logical order you could follow throughout the manuscript.
Line 29: Which seasonal changes? Meteorological changes?
Line 31: By ‘groundwater age’ you mean what exactly? The residence time of the groundwater, the travel time of the groundwater to the stream? The age of the fraction of streamflow fed by groundwater?

Introduction
Line 41-45: Here you introduce the terms ‘age’, ‘residence time’ and ‘travel time’ somewhat interchangeably and without proper explanation. Also, I would mention flow pathways first (they cause/control the travel times and not the other way around). For example: ‘The pathways of water through the terrestrial landscape to stream networks and associated age is a…’.
Line 55: Only if you assume that the groundwater is fully mixed. If not, even groundwater can still be associated with certain hydrologic events.
Line 56-57: You should at least mention that a large fraction of young stream water can also be derived from shallow subsurface flow in the soil or at the soil bedrock interface (some call this ‘interflow’).
Line 65: It is the dampening of the isotopic tracer signal that provides MTT estimates.
Line 68: Precipitation-discharge relationships are not similar to the mentioned gamma distribution transfer functions. I would leave them out of this argument.
Line 86: How are ‘isotope models’ a method to calculate travel times? I understand how particle tracking can be used to calculate TTDs but ‘isotope models’?
Maybe you are talking about models that use solute transport routines (assuming conservative non-reactive tracers) to trace water pulses through a catchment system? Fitting references would include Remondi et al. (2018) and Heidbüchel et al. (2020).
Line 117: Age distributions or travel time distributions? Please be consistent in your terminology.
Line 118: Seasonality in what? Temperature? Precipitation amount? Weather in general?
Line 119: How do you compare calculated travel times to 10-year seasonal isotope signatures? Do you compare the relative damping of the signatures for different years to the calculated travel times?

Method
Line 169: What is a ‘rapid transition in hydrology’?
Line 177: ‘measurements’ instead of ‘results’.
Line 177-179: Comprehensibility would benefit enormously from a more concise writing style. This sentence is a good example:
Instead of writing a long and convoluted sentence like this one:
‘Some of the sub-catchments are affected by evaporation from lake surfaces that result in isotopic fractionation that, consequently, affected the signal (Leach and Laudon, 2019)’.
Better write:
‘Isotopic fractionation caused by lake surface evaporation affects the isotopic signal of some of the sub-catchments (Leach and Laudon, 2019).
Line 179: ‘…was corrected by accounting for the percentage…’
Line 183: This is not a logical conclusion. Just because there is no groundwater recharge during winter this does not mean automatically that the stream is only fed by groundwater only. It could still be fed by overland flow from meltwater or from rain on snow...
Line 184: What aspect of the isotope (measurements? time series?) should be statistically related to average age of the groundwater?
Line 184: Are you aware of the fact that the average age of the groundwater is the mean residence time and not the mean travel time of the water reaching the stream? They can be very different. You need to be very careful with the terminology you are using. Please go through the manuscript and check each occurrence of the words ‘age’, ‘MTT’, ‘travel time’, ‘residence time’. Don’t use two terms for one and the same concept – like ‘average streamflow age’ and ‘MTT’ or ‘average groundwater age’ and ‘mean residence time’.
Line 185: Up until the point where full mixing is reached. Full mixing of what?
Line 187: But now imagine a year where the average input signal is equal to the average annual precipitation signal (that can also be found in the deeper groundwater). In this case you cannot draw a conclusion about how old the winter groundwater signal is because it will be close to the long-term average no matter if there is more young recently recharged or more old groundwater in the mix. Maybe you can use the variability in the winter stream isotope signal. But I would like to see a better explanation on why this method would work.
Line 191: But does the preceding spring not influence the summer signature also? So how can you assume that the difference between winter and summer signature gives you the young water fraction in the summer if there is another overlaying signal from the spring? Somehow you have to take this into account.
Line 253: Why are the UZ processes only active in the SZ? Should they not be active first and foremost in the UZ?
Line 259: Do you mean all ‘soil layers’ not deeper than 2.5 m below the ground surface being average into one soil type?
Line 259-265: You used ‘In the Krycklan model…’ at the start of sentences three times in this short paragraph.
Line 276: Streamflow accumulated of what period of time? It makes a big difference if it is accumulated over a day or a year. So I would write ‘…able to reproduce daily accumulated discharge…’.
Line 306: Numerical constraints restricted the number of particles released to 0.5 particles/10 mm modelled groundwater recharge PER GRID CELL, which corresponds to a total of approximately 0.6 million particles FOR THE ENTIRE MODELED AREA IN THE FIRST YEAR.
Line 318: However, that does not mean that the distribution is significantly skewed if the SD is larger than half of the average.
I think what you want to express is that if the SD is smaller than half of the average, the distribution is not significantly skewed.
Line 324: Good job explaining why you chose to use the geometric mean. Makes sense to me.

Result
Line 354-355: Be more consistent in the way you describe the catchments (size, soil, landuse). For example you say that C20 and C16 are both silt-dominated, yet in the parentheses you only repeat this for C20.
Line 371: However, the groundwater MTT of mire-dominated catchments supposedly showed less variation due to the smaller amount of groundwater recharge in your (not-overland-flow-corrected) particle tracking model. Is that the case?
Line 393: How are the annual and seasonal MTTgeo values computed? Are they flow weighted or simple time averages?

Discussion
Line 445: What do you mean by ‘gamma transformation method’?
Line 453: Do you show this scaling function anywhere? Would be good to see it written out.
Line 483-484: These theoretical consideration strengthen the results of a winter MTTgeo being LONGER OR EQUAL TO four years. Are they also telling you that winter MTTgeo is BETWEEN four AND SIX years?
Line 529: What exactly do you mean by ‘synchronicity’? The seasonal patterns? Do you mean that the changes occur at exactly the same time during a year?
Line 531: ‘the areal coverage of silt’? Do you mean ‘the spatially averaged silt content of the soil layer’? Or ‘the areal fraction of soils with a certain silt content’?
Line 535: The variability of what?
Line 532-543: I would also mention the correlation between catchment size and silt content you observed when discussing this issue.
Line 545: ‘Silt fraction effect’ is unprecise wording. What fraction? In the soil? Across the landscape?
Line 570: What do you mean by ‘…silt dominated areas, that have more consistent hydrological conductive with soil depth…’?

Conclusions
Line 583-585: This is interesting and you should explain what this would entail. Where and in which way would transit times change in a warmer climate?

Figures:
Figure 1: In the caption you use strange sentences (‘(b) The soil map used in the Mike SHE flow model and is based on the soil map…’, ‘(c) Depth to bedrock from the Swedish Geological Survey (2014) is shown in meters…’).
Figure 3: There are no labels on the graphic. Where is the UZ, the SZ? I don’t see the ten calculation layers either.
Figure 5: Are these TTDs specific for a certain point in time? Are they forward or backward TTDs? Are they weighted averages of multiple TTDs (master distributions)?
Figure 6: The blue outline of the surface runoff gives the impression that there is surface runoff happening constantly in b), c) and d). Rather make it a blue fill without an outline (like the green fill for the groundwater contributions).
Figure 7: I recommend using the same scale for MTT, BC and YWF for all axes. This way the different relationships (as well as the season-specific variabilities) become much clearer.
In the caption better write something along the lines of ‘Relationships of seasonal MTTgeo and young water fractions with seasonal stream isotopic composition and base cation concentration’.

Tables
Table 1 and 2: Just a suggestion: Why not grouping the catchments and nested sub-catchments instead of listing by catchment number (so all the green catchments together and all the red and blue catchments together as well)? The numbering is somehow arbitrary.
Table 5: I would rearrange the order in all the tables and group the catchments as mentioned earlier.
In the caption rather write ‘…that is younger than three months…’ instead of ‘…that is less than three months…’.
Table 6: Why did you remove the values that are smaller than 0.5? Please add them again - in the caption you are still mentioning the darker colors. Otherwise I like the new design (just make sure you either close all or none of the boxes around the seasons).

Appendix
Line 869-870: ‘whereas’?

References
Remondi, F., Kirchner, J. W., Burlando, P., & Fatichi, S. (2018). Water flux tracking with a distributed hydrological model to quantify controls on the spatio‐temporal variability of transit time distributions. Water Resources Research, 54(4), 3081-3099.
Heidbüchel, I., Yang, J., Musolff, A., Troch, P., Ferré, T., & Fleckenstein, J. H. (2020). On the shape of forward transit time distributions in low-order catchments. Hydrology and Earth System Sciences, 24(6), 2895-2920.

Hide

ED: Publish subject to minor revisions (review by editor) (01 Feb 2021) by Conrad Jackisch

AR by Elin Jutebring Sterte on behalf of the Authors (10 Feb 2021) Author's response Author's tracked changes Manuscript

ED: Publish subject to minor revisions (review by editor) (02 Mar 2021) by Conrad Jackisch

Dear Elin and co-workers,

Congratulations for your last revisions. Thank you for all the efforts which went into your publication in our special issue. I have to apologise for the delay in handling your latest version, which was mainly due to my own uncertainty how to proceed. While in my view your paper is ready for publication in general, I came across a couple of unnecessarily sloppy details, which might be easy to resolve. Moreover, I am under the impression that your publication could be a very nice example for transparent science including the fundamental analysis code.
Hence I decided to ask you once more for a quick feedback in a round of minor revisions before final acceptance.

First of all the data and code: Your statement links to the Krycklan data website, which is fine. I made the effort to download the data and do some very quick plotting. What I could not find is data about the isotopic concentration in the precipitation. Further the adjusted 18O observations appear to be lacking. Moreover, I do not see how I could possibly rework your results when starting MikeSHE from scratch and without the geo data. Since it is the year 2021 and since our special issue is also about transparency in data and code, I would like to persuade you to scope your options to publish some sort of repository alongside your paper. I simply see that your study is a very nice example of TTD and MTT analysis and people will hook to your work much more easily, when it is possible to really follow your calculations. I fully understand that the data policy of the Krycklan data explicitly seeks to avoid a copy of this data. This is fine. But for your further data and/or model setup/model results this could be easily feasible through github or any other DOI providing repository.

Now my minor suggestions along the last review of your manuscript:
Abstract:
Why is the “Krycklan” catchment name avoided until L112? I suggest to include it in the abstract near L25, since it is a well known experimental catchment. This could also strengthen the reference of the third sentence. It will also help meta analyses.
L31ff. The sentence was modified and is fragmented now. Maybe: Catchments with mixed soil landscape settings typically displayed larger variability in seasonal MTTgeo, as contrasting hydrological responses between different soils (e.g., mires, till and silty sediments) are integrated. One more detail: I am not sure if mires and till can count as “soil types” (which may be a matter of languages used in the different classifications). Maybe simply “soils” would solve the issue?

Main Document:
L135f. Maybe: the characteristic VEGETATION of this boreal landscape are Scots pine and Norway spruce… I would not see the vegetation as characteristic features.
L141 Why is the hydraulic conductivity given here? 5e-5 m/s is not coherent with table 3. Moreover it is difficult to refer to, when the other soil landscape characteristics are left for sec 2.3. Further a decreasing hydraulic conductivity is nothing specifically rare… I suggest to check again, what information is really characterising the different soil landscapes.
Fig. 2: I still struggle to read this conceptual figure – not because it was too complex but rather not really in touch with the data. As mentioned, I tried to catch up with your approach by downloading some data from the database. I see your point in using the difference from the ∂18O winter. But what I could not really trace back is if the winter streamflow covers a range and the difference is calculated to either upper and lower bounds of this range, as the figure suggests. Your text does refer to the averages without any distribution. When checking in with the data, I could see the spring event as very pronounced signal while the summer is far less depicted. I was wondering if this figure might benefit from using real data or at least a curve for an idealised streamflow? Could you please clarify the winter streamflow range reference? Did I understand correctly that the isotopic concentration “raw data” from the Krycklan data base was adjusted (L179) and that the final merged product of 18O concentrations was used as reported in Table 2?
Table 2: I could not see any isotopic concentration records of the precipitation in the Krycklan database.
L285f. Did I understand correctly that until here, MikeSHE was setup based on common rainfall, runoff and groundwater dynamics but without information about the isotopic concentration dynamics or BC? Maybe referring to “all available data” is sparking some confusion?
Fig. 7: Here the concept explained in Fig. 2 is applied to evaluate the model performance. Referring to the last comment, 18O and BC are only used as evaluation reference, correct? I suggest to either include the ∆ in the subfigure titles in c and e. Or even more easy to digest why not printing ∂18O Winter – ∂18O Summer?
L476f.: Did I understand correctly that MikeSHE simply adds a particle for every 20 mm recharge but that the isotopic concentration dynamics of the precipitation was not used? Does this imply that the percolation is actually neglected? In addition I find your argument slightly self-fulfilling: Since you focus your analysis on groundwater dynamics you can only retrieve dominance of the groundwater domain. All interpretation whether percolation and unsaturated zone water is negligible or not is actually just reiterating your assumption. I think this could be formulated more clearly here.
L542: I suspect “this" refers to the finding about catchment size. Maybe something like this is more clear? “However, a correlation of travel times to catchment size appears to be a spurious relationship.”
L602: why is there a dot after IPCC?

I hope you find my comments helpful to fortify your paper and its outreach. I would be very happy if you can find a way to provide a more specific and complete reference to the data and at best even a condensate of your code.
All the best. Thank you again.
Conrad

Hide

AR by Elin Jutebring Sterte on behalf of the Authors (05 Mar 2021) Author's response Author's tracked changes Manuscript

ED: Publish as is (15 Mar 2021) by Conrad Jackisch

AR by Elin Jutebring Sterte on behalf of the Authors (16 Mar 2021)

Short summary

A numerical model was used to estimate annual and seasonal mean travel times across 14 long-term nested monitored catchments in the boreal region. The estimated travel times and young water fractions were consistent with observed variations of base cation concentration and stable water isotopes, δ¹⁸O. Soil type was the most important factor regulating the variation in mean travel times among sub-catchments, while the areal coverage of mires increased the young water fraction.