Overview: The authors presented a very interesting experimental design to estimate the radial velocity distribution of xylem water. A sample of a tree was cut, a dye boundary condition was introduced at one end and allowed to percolate through the sample. Once dye was visible on the downstream end, the sample was divided into sections. Image processing software was used to identify areas of dye to determine what xylem was active along the length of the sample. The study carries implications for both field sampling plants for isotopic analysis and numerical modeling of plant water uptake and storage. Overall, I think the idea is novel and highlights a problem that has been largely overlooked in field sampling campaigns. I enjoyed reading the paper and think this is a strong contribution. I believe the experimental design and the writing could be slightly improved. Please see comments below.

I will also disclose that I am referencing two of my own papers in this review. I’m referencing one paper because the authors comment on it several times. I’m referencing another because it’s one of the few published applications of SAS functions to modeling transport velocities within plants. I strong emphasize that I am not expecting citation of work that I am affiliated with.

Major comment: The proposed methodology hasn’t been validated with a parallel established method. 

There is some discussion of how these results compare to radial variations in xylem water velocity measured with sapflux techniques; however, these prior measurements were made at different heights, depths into the xylem, and on different species. This is a concern to me particularly because the proposed methodology required subjective visual tuning of the NDBI classification of dye versus no-dye in each cross section. This visual tuning seems to have resulted in an underestimate of the dye-stained area (Fig. 2). I would also guess that this underestimate is what causes the authors to calculate that a substantial proportion of outer xylem tissue does not conduct water (Fig. 6c-d). It may be the case, but this type of finding would impact the conclusions of many prior studies and should therefore be carefully validated.

The ideal resolution to this comment would be to provide a validation or comparison of the technique. I would imagine that this experiment could accommodate sapflux sensors installed along the sample and/or xylem water isotopic sampling (if labeled and dyed water were used at the boundary condition). I recognize that this would be a substantial amount of work. At the very least, I strongly caution against using the results of this preliminary experiment to guide field sampling and model development and think the results and conclusions could be tempered a bit more than they already have been.

Specific comments:

Section 1: I appreciate the focus on process-based modeling of root to xylem conduit time lags. I would also point out that research using storage selection (SAS) functions have been presented as a parallel way of dealing with this problem. These models don’t need to explicitly consider where water spent time (i.e., inside the plant versus in the soil). SAS functions can numerically represent complex age-distributions of water in xylem. I understand that this isn’t the primary focus of the paper, but this comes to mind as I read the argument the authors are making. Here are several examples:

Evaristo, J., Kim, M., van Haren, J., Pangle, L. A., Harman, C. J., Troch, P. A., & McDonnell, J. J. (2019). Characterizing the fluxes and age distribution of soil water, plant water, and deep percolation in a model tropical ecosystem. Water Resources Research, 55(4), 3307-3327.

Knighton, J., Souter‐Kline, V., Volkmann, T., Troch, P. A., Kim, M., Harman, C. J., ... & Walter, M. T. (2019). Seasonal and topographic variations in ecohydrological separation within a small, temperate, snow‐influenced catchment. Water Resources Research, 55(8), 6417-6435.

Smith, A. A., Tetzlaff, D., & Soulsby, C. (2020). Using storage selection functions to assess mixing patterns and water ages of soil water, evaporation and transpiration. Advances in Water Resources, 141, 103586.

Line 49: The process-based modeling papers described by the authors all shared a similar challenge in that they were attempting to simultaneously simulate the both the soil water balance (the root boundary condition) and xylem water isotopic observations. Some of the model error in representing xylem water was likely error that cascaded down from an imperfect representation of the soil water, and not only an imperfect representation of xylem water transport. Smith et al (2022) demonstrated this directly. This might be worth mentioning.

Line 55: This transition in text is kind of abrupt. The text goes from advances in modeling the soil-plant system to measurement techniques. Would this paragraph make more sense moved down to be the first paragraph of Section 1.2?

Line 75: A definition of “mobile” would be useful here. Is “mobile” water extracted under a certain pressure? Many researchers remove heartwood from tree core samples prior to isotopic analysis to sample only the “mobile” water; however, recent studies have suggested that heartwood is “mobile” to a certain degree. I think the field at large needs a better definition for “mobile.” For example, see Fabiani et al (2022).

Fabiani, G., Penna, D., Barbeta, A., & Klaus, J. (2022). Sapwood and heartwood are not isolated compartments: Consequences for isotope ecohydrology. Ecohydrology, 15(8), e2478.

Line 87: Possibly no one has done this with stable isotopic techniques or dye tracers, but there are many studies where the radial variations in xylem velocity have been observed (Ford et al 2004). Maybe I’m not understanding what the authors are suggesting, but this seems a like a new method of measuring something that has been measured many times before.

Ford, C. R., McGuire, M. A., Mitchell, R. J., & Teskey, R. O. (2004). Assessing variation in the radial profile of sap flux density in Pinus species and its effect on daily water use. Tree physiology, 24(3), 241-249.

Line 107: How long did this dye breakthrough require?

Figure 6c-e: This is implying that some shallow xylem tissue existed that was not transporting water. It seems more likely that this was because the NDBI algorithm was missing areas where dye was showing up weakly. You can see that this is likely the case on the left-hand side of Figure 2a and 2c. There is a clear blue-green coloring along the left of the cross section that was not identified by the algorithm and (I’m assuming) coded as “immobile.” The authors also point out that the blue dye leaked from the last cross section (mobile) but did not stain this cross section (not identified as mobile). I would guess that the reported “immobile” water percentages are more related to methodological limitations than evidence that immobile water is being sampled in short cores.

Line 225: I’m not sure if it would work with this the methodology, but we often apply mineral oil to increase visibility when counting rings on tree cores. It makes the wood grain much more distinct. This might be an inexpensive way to quickly improve image quality.

Line 230: You could also include a reference tile or swatch next to each stem cross section for image post-processing to a visual standard.

Line 284: The wording here is a bit awkward: “a lot more, very common trees.”

Line 295 - 296: This recommendation seems like an overreach given that this methodology hasn’t been validated (see major comments).

Line 304: “wide range of velocities” maybe isn’t the most accurate wording. As discussed in section 4.2 this study shouldn’t be taken as a measurement of velocity because of the unusual boundary conditions. Maybe instead: “a large variance in the radial distribution of velocities” which avoids talking about the actual velocities.

Line 305: Minor comment here, but Knighton et al (2020) investigated both piston flow and a fully-mixed reservoir, which was a little more complicated than piston flow. 

Line 305: There is also a possible difference in the temporal scales between that modeling experiment and this study presented in the manuscript. I’m assuming this study occurred over a time period of less than one working day (<8 hours). If a boundary condition is introduced and held constant for a sufficient period of time (e.g., rainfall onto soils followed by a period of no rain for a week) the radial distribution of xylem velocities will stop mattering after the slowest path has reached the sampling height. The variation in flowpath velocities would only be a strong consideration if the water boundary condition at the roots was rapidly changing relative to the transit time between roots to the sampling point, which might not always be a real-world concern.

Line 316: Something wrong with this reference “?Dubbert”

- James Knighton

General comments:

The study of Seeger & Weiler assesses the distribution of water transport velocities within a tree stem, using a dye tracer and a virtual experiment. The study’s main goal is to test the assumption δRWU = δXylem, often used in isotope-based plant water uptake studies. Moreover, it studies how different xylem sampling approaches affect derived water transport velocities. The study finally concludes that δRWU = δXylem is too simplified and that the variation of water transport velocities needs to be better addressed in isotope-based plant water uptake models and when sampling xylem water ofr isotopic analysis.

The major constrain of the study is that it was not directly compared to, for instance, sap flow or water stable isotope measurements. Also, looking at the cut surfaces of the stem, the cutting seems not to have always worked so well (Fig. 2a) and probably has made the image processing more difficult and less accurate. Yet, the study is a nice starting point for further studies on that topic and it provides good suggestions for further improvements of the method. The ideas of this study is original, and the study will be a great add-on to the current literature.

Specific comments:

The table and figures are lacking important information. Explanation for abbreviations is widely missing. All figures need to be explained better in the figure legends, not just in the text.  A lot of information can be only found in the text, see e.g. information in lines L189, L209-210. The legend of Fig. 4b for instance is not referring at all to the flow path length (x-axis of the figure). 

It would be great if the figures would also use/refer to the abbreviations from the equations for better understanding.

The connection between 1 Introduction, 1.1 and 1.2 should be made clearer. The relevance of 1.1 and 1.2 only get clearer when the objectives (1.3) are listed. Some sentences at the end of section 1 should introduce 1.1 and 1.2.   

How is the variation within a xylem class? Looking at Figure 4, also the trunk side of the sampled tree mattered. Figure 3a assumes no large variation within the xylem class? This point of variation within a xylem depth class should be also discussed.

Will you share your codes with the community?

Line by line comments (mainly minor):

L3: you only study the plant water uptake/transport starting from the stem base

L8: "sampling approaches for xylem water stable isotopes"

L12: since they (?, the isotopic signatures?)

L17:  water stable isotopes

L21: maybe rather introduce abbreviation δsoil here

L26: "soil water isotopic data"

L29: you could add the information in the brackets directly in the sentence, nicer to read

L45: led

L47: phrasing, better: „were not able to“

L49: "of an in-situ"

L78: for consistency: "all water that is contained"

L79: CVD abbreviation not introduced

L82: the borehole method likely only sees the last few cm of the borehole before the airstream is carried out of the stream

L95: it would be enough/nicer just to write: a dye tracer experiment (Fig. 1). We… 

L96: nicer to read if you just add the info in the brackets to the text, for example, „segment of a 3.5 cm“

L98-99: add maybe „for this study“ or “here”

L107: add company information etc. for brilliant blue

L115: Figure 2, could you add a scale as a reference?

L118: „where we used“ 

L118: how big is one pixel?

L120: how were the classes defined exactly? Be more precise. Why six classes? I would already add “(Fig.2c)” here

L124: might be nicer to write „Eq. 1“ etc. As you also reference to them later.

L133: „required velocities u_n to reach“

L152: add more information „does not matter because…“

Figure 3; four different kinds of tiny circles? do you mean shapes/colour groups?

L160-165: explain / reference the choice for innermost=0, outermost decreases again

L170: “any study”

L172: „different xylem sampling approaches will capture“

L174. intercomparison, inter-comparison reads nicer

L185: how much was the thicker end of the stem?

L186: you could add „Eq. 6“

L194: „virtually“ is confusing here as this is the actual experiment

L199: equation 2.1.2??

Fig.5 legend. „per xylem depth class“, (c) is actually the xylem radius (be more consistent)

Fig.6 Frequency: y-axis labels

L230: led

L238: led

L247-251: maybe make two sentences out of this very long one

L252: no comma

L254: „as fast as“

L255-L261: mention that velocities can be highly species-specific and size of tree etc.

L258: wasn’t this even in three heights?

L265: did you sample your stem for isotopic analysis? so that you could compare it directly?

L266-274: these are all different species, might be worth to emphasize this more

L270: start new sentence after “… beech)”

L278: led

L284: phrasing

L285: „more conductive pores“ than? be more precise

L287: „trees. It is“

L288: behave differently

L294: overrepresentation, over-representation reads nicer

L316: ?Dubbert, delete“?“

L316: phrasing