Review of Marshall et al., for HESS

This manuscript describes a study in which a deuterium tracer was added to soil during drought, allowing for the opportunity to see the timing and spatial pattern of trees using that added water. Continuous isotopic monitoring capabilities were crucial to observing the breakthrough of that tracer in a set of trees. Their differential distances from the location where the tracer was added demonstrated a) the relative timing of later transport from through roots to each tree stem and b) the likelihood that numerous trees’ roots were taking up water from the 1m2 location of tracer addition.

The results are simple and straightforward (with a few caveats described later), but the ability to conduct an experiment like this is a complex challenge. Consequently, it is a rare to have insights into a process such as this, and so I think this study will represent a valuable contribution to the field. While I find the discussion of the results to be framed in a way more appropriate to an ecophysiology audience than a hydrology audience, this work still fits within the scope of HESS. I enjoyed the writing style and I especially appreciate the blunt description of the challenges and mysteries; this sets a positive example that I would like to see more scientists follow.

I think the greatest area for improvement would be to follow the suggestions of the first posted review, regarding more in-depth consideration of how to address mass conservation issues. I do not have specific suggestions beyond those.  

While I know the borehole method has been tested and described in other papers, my concerns with the method pertain to the hole implying a severing of the xylem transport pathway. Is it capturing water in transit up the tree, or some signal that has diffused from active xylem into this less active wood storage (after being drilled)? By focusing on the tracer arrival timing, interpretations seem robust to this issue. However, it could explain the mass conservation failure. Can you elaborate on this further based on your experience in experimenting with this method?

71-72 The need for quantifying overlap is mentioned here. I’d like to see elaboration later in the discussion, within the context of this ownership hypothesis.

Figure 1 might benefit from having the scales in meters rather than degrees lat / lon.

376-378 Inclusion of tree metrics would benefit the contextualization of findings. What are the crown sizes? Perhaps this could be shown in Figure 1.

423 Can you clearly state the “ownership hypothesis”. If it’s simply that there exists “a continuum between complete domination of uptake by a single tree, analogous to owned and defended property, vs. complete sharing of the benefit”, I do not see how this is a useful hypothesis. It seems obvious that this continuum would exist and it is unclear how this hypothesis would be tested. I would hope to see a hypothesis rooted in what would drive the owning-to-sharing variation.  

429-439 (including Figure 7): More elaboration would be useful here. I struggled with precisely understanding the point.  

Colleagues,

I present this review as a copy-paste of the one I posted erroneously as a general public comment. I was recruited by the AE to provide the review originally and the posting as a public comment was my error. I re-post here to satisfy the journal requirements and reduce stress on the reviewer pool.

--RFK

---

This manuscript presents results of a field experiment to identify root water uptake using isotopically labelled irrigation. The measurement techniques are state of the art, which is to say that they are difficult to implement and interpret. The results are satisfyingly clear in terms of temporal behavior of isotopic composition of soil water and sap, and the experiment is thus a rare opportunity to quantify processes that are difficult to observe. The strength of the results lies in the spatial observations of tree water uptake, and the weakness is the mass conservation estimates. Some improvements in mass conservation estimation might reduce some of the considerable uncertainty that derives from recovering less than 5% of the tracer.

L162 “UD” does not appear in Eq 1, and is later defined as something else (L167). Please move L168 to L162.

L171 what is the exact scope of this non-evaporation assumption? For example, do the berry bushes count in the reported LAI or not? Also, mosses are a nontrivial pathway for E (e.g., Heijmans et al. Global Biogeochem Cycles 2004). Finally, Fig 4b, 5a pre-labeling lce<0 and vertical profiles of 2H both suggest soil evaporation (though there are also other possible explanations for these conditions). If the flux tower data and pre-labeling soil isotopic profiles are not sufficient to resolve the soil/moss/understory(?) E flux component, some rough, literature-based or isotope-based estimates would be helpful for estimating biases introduced by assuming it to be zero.

What is the likely magnitude of downward unsaturated flux—i.e., to a zone below the soil moisture sensors? If unaccounted for, this flux would give upward bias to estimates of water consumption in upper layers—more bias the shallower the layer.

Fig 1 shows 8 trees within 4 m of the injection. This is a tree-dense part of the forest. What does this mean for extrapolation of the results to the stand scale?

L198 the whole-tree-bore equilibration method causes samples to equilibrate with all water along the hole, whereas the tree water might be dominated by recent uptake only at the outer few mm of xylem. Just as Marshall et al (2000) addressed this, it would be helpful to add some discussion of the likely magnitude of this effect in this experiment.

I think the Label Recovery section of the discussion more properly belongs in the results. Also, what are the consequences for the soil depth uptake calculations of accepting some alternative hypotheses about the fate of the unrecovered 95%+ of tracer?

Fig 6 and its caption could be improved to explain the meaning of “days after labelling”—day of first observation? Also, isn’t it more logical to swap the axes?

Are there no measurements of the monitored trees available? I expect differences in rooting and uptake might be related to things like relative crown size and sapwood area.