Effect of parameter choice in root water uptake models &ndash; the arrangement of root hydraulic properties within the root architecture affects dynamics and efficiency of root water uptake

Bechmann, M.; Schneider, C.; Carminati, A.; Vetterlein, D.; Attinger, S.; Hildebrandt, A.

doi:https://doi.org/10.5194/hess-18-4189-2014

Articles | Volume 18, issue 10

https://doi.org/10.5194/hess-18-4189-2014

© Author(s) 2014. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/hess-18-4189-2014

© Author(s) 2014. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 18, issue 10

Research article

|

27 Oct 2014

Research article |

| 27 Oct 2014

Effect of parameter choice in root water uptake models – the arrangement of root hydraulic properties within the root architecture affects dynamics and efficiency of root water uptake

M. Bechmann, C. Schneider, A. Carminati, D. Vetterlein, S. Attinger, and A. Hildebrandt

Download

Final revised paper (published on 27 Oct 2014)
Supplement to the final revised paper
Preprint (discussion started on 15 Jan 2014)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC C40: 'Referee comments on HESSD 11, 757, 2014', Anonymous Referee #1, 31 Jan 2014
- SC C145: 'Note to anonymous reviewer #1's comment', Marcel Bechmann, 20 Feb 2014
  - SC C149: 'Figures in a seperate supplementary', Marcel Bechmann, 20 Feb 2014
    - RC C204: 'Answer from referee # 1 to Bechmann et al.', Anonymous Referee #1, 27 Feb 2014
      - RC C218: 'Correction', Anonymous Referee #1, 28 Feb 2014
      - SC C549: 'Rreply to anonymous reviewer #1's answer', Marcel Bechmann, 20 Mar 2014
- AC C1345: 'Final response to anonymous reviewer #1', Marcel Bechmann, 12 May 2014
EC C118: 'Invitation to reply to Reviewer's comment', Nadia Ursino, 12 Feb 2014
RC C270: 'Comment on "Parameterizing complex root water uptake models– the arrangement of root hydraulic properties within the root architecture affects dynamics and efficiency of root water uptake" by M. Bechmann et al.', Anonymous Referee #2, 04 Mar 2014
- SC C633: 'Reply to reviewer \#2, referee comment RC C270', Marcel Bechmann, 24 Mar 2014
- AC C1347: 'Final response to anonymous reviewer #2', Marcel Bechmann, 12 May 2014
EC C272: 'major concerns raised by the reviewers', Nadia Ursino, 04 Mar 2014

Peer-review completion

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

ED: Reconsider after major revisions (17 May 2014) by Nadia Ursino

Dear Marcel and Coauthors,
Your paper stimulated the discussion of a number of outstanding topics. I appreciated the effort that the Reviewers and you put into the debate.
I decided to reconsider your paper after major revision and I would send the revised version to at least two independent Reviewers in order to clarify if the major disagreement with AR1 stems from a different interpretation of basic principles or if there are fundamental assumptions that seriously prejudice the results. My decision relies on the reviewers’ judgment and on the arguments raised on both sides. I cannot ignore the major concerns on the scientific significance of your results raised by AR1.
There is no guarantee that the paper will be accepted after this second run of review.
I believe that addressing the reviewers’ comments in a revised version would add considerable value to your work and eventually make it suited for publication in HESS, and I hope that every conceptual problem could be settled, as well as the controversial interpretation of relevant physics.

Please find below a list of the major points that I expect to be clarified in the revised paper.
THE HYDRODYNAMICS OF ROOT UPTAKE
The constant value of transpiration, the absence of redistribution, a schematic representation of root architecture, soil and soil moisture homogeneity lessen the correspondence between your conceptual model and reality. Root architecture determines the connection between the plant and different soil depths, where the boundary conditions for the root uptake dynamics may be very different (as pointed by AR2). To which extent, focusing on root resistance only may misrepresent the real world water use efficiency, should be stated clearly.
THE DEFINITION OF INDEXES
According to your reply to AR1 the interrelation between “effort” and “yield” should be declared. The disagreement with AR1 seems to originate from previous literature definition/interpretation of the concepts behind indexes. I think that the reader deserves to know what is new in your definition of yield, why you need to introduce effort instead, if there is a correlation between the two indexes (as stated by AR1) and to which extent this correlation is due to your simplifying modeling assumption (constant value of transpiration, the absence of redistribution…).
ACCURACY OF RESULTS
The link between numerical discretization and model outcome must be clarified according to your reply to AR1.
MAIN CONCLUSIONS AND LINK TO REALITY
The index analysis based on the concepts of axial and radial limitation, suggests that an optimum root development strategy could exist. As suggested by AR2 you could point out if there is any experimental evidence that corroborates your findings or even just the basic assumptions in your conceptual model.

Hide

AR by Marcel Bechmann on behalf of the Authors (25 Jun 2014) Author's response Manuscript

ED: Referee Nomination & Report Request started (27 Jun 2014) by Nadia Ursino

RR by Anonymous Referee #3 (30 Jun 2014)

RR by Anonymous Referee #1 (08 Jul 2014)

Suggestions for revision or reasons for rejection

General comments:

This manuscript raises interesting points, among which the impact of root topology and maturity on plant water status and root water uptake dynamics. The text is well structured and figures and tables are clear.

The improvements of the manuscript as compared to the previous version are significant, among which:

- The differentiation of the existing concept of “hydraulic isolation” (i.e. with increasing axial resistance, less water is absorbed in the distal part of a root) and the newly introduced concept of “axial limitation” (i.e. root hydraulic resistance would increase with root length at some point).

- General definitions of the newly introduced indexes “effort” and “water yield”.

- The clarification that effort is not a measure of plant hydraulic resistance, even though it is sensitive to it.

- Variables are also clearly declared with their units.

However, the theoretical background is definitely subject to the artefact mentioned in the first part of the reviewing process. I believe that it is misleading and should be removed from the manuscript. The study on (un-)branched roots is seemingly still slightly affected by the artefact for young roots, but especially the absence of vertical soil water flow coupled to the non-equilibrium initial soil water condition might be responsible for the increased effort for longer roots. If not, it should be clarified whether or not the increased effort actually results from an increased (absolute) xylem water potential.

Specific comments:

RC1, P4, L14: Lobet et al. [2011] and Lobet and Draye [2013] also did excellent work on the subject recently.

RC2, P4, L24: Doussan et al. [2003] does not appear in the bibliography. Isn’t it 2006?

RC3, P6, L2: Here it is mentioned that the soil water potential is initially homogeneous. However it is further explained that the initial water content is initially homogeneous. Was it first meant that the soil matric potential is initially homogeneous? Is the gravitational potential accounted for? This point is important for what follows (RC13).

RC4, P6-7, L24-11: Equation (1) corresponds to the case of a root composed of a single segment. As demonstrated by Referee #1 of the first reviewing process and by the authors in the supplemental figure S1(a), the estimation of total root resistance (Rtot) using a single long segment is inaccurate and leads to overestimated absolute values of collar water potential.

About Eqs. (2-5), an aggregated radial resistance is valid if xylem potentials are uniform (assumed on short distances -> short segments), while an aggregated axial resistance makes sense if no radial conductances generate parallel pathways in the circuit.

Equations (6-7) and the predicted minima are thus incorrect for hydraulic architectures with significant axial resistances and radial conductances.

RC5, P7, L23: The term nicely does not sound very scientific…

RC6, P7, L25-27: Here is explained that the indices account for both root and soil hydraulics. The indexes are indeed sensitive to both root and soil hydraulic resistances (as long as both are present). However, only root hydraulic resistances are accounted for in most of this study (all of it except simulations with aRoot).

RC7, P9, L4-6: The supplementary mentioned here demonstrates that for a mature root, a discretization of 100 segments is largely enough to avoid any significant artefact on the estimated total root resistance for a mature root. However, this was not demonstrated for young roots, which are more sensitive to the artefact due to their limited axial conductance. In the lower-left subplot of the attached figure, it is visible that an error of a few percents remains for long young roots. However this small error alone probably does not explain the significant change of effort visible in Fig. 3 of the manuscript (other possible reasons are discussed in RC13). Actually, a simple way to avoid the artefact of “increasing root resistance with root length” is to model growth by adding root segments of constant size instead of elongating existing segments (see low-right subplot of the attached figure).

RC8, P9, L18-20: I agree with the authors, the simplifying assumptions chosen in this part of the manuscript are good as long as they do not lead to misleading conclusions. Their goal is to emphasize how simple root topologies and hydraulic properties affect the newly introduced indices, before testing these two indices on more complex hydraulic architectures.

RC9, P13-15, L20-8: Here the authors provide clear definitions of water yield and effort, which I think was very important in order to correctly interpret the results. The following minor suggestions are rather a matter of taste:

In the definition of water yield, the authors selectively exclude water transpired during water stress. This justifies the fact that water yield does not change once water stress was reached in the simulations with constant Tpot, and that it continues to evolve in simulations with day-night fluctuations of Tpot. I found it counter-intuitive that water transpired during stress events is not part of the yield; after all, the index including stressed transpiration would discriminate fully transpiring plants from plants under water stress anyway (even though not as much). It might actually be problematic for the application of the index on real plants, for which it is not always easy to delimit when stress begins and ends, while for a plant in pot the measure of total cumulative transpiration by weighing (not discounting stressed transpiration) is easy.

Water yield and effort are said to represent the benefit and cost in the quantification of root water uptake efficiency. I easily see cumulative transpiration as a benefit (as it can be translated in terms of cumulative photosynthesis product) and leaf energy level and root length as costs (as a very low xylem water potential implies consequences that might be bad for the plant, while building roots involves a cost in terms of carbon and other elements). However the fact that water yield has root length (a cost) as denominator does not make the classification in terms of cost and benefit as straightforward. As is, water yield represents some kind of average benefit per root length, which does not make it as clear whether it is beneficial for the plant to build more roots. Actually increasing the root length might very well delay water stress (beneficial) but decrease water yield. Also, in effort, the energy cost is divided by the cumulated transpiration, the latter being a benefit. I believe that both indices are efficiency-related, but should not be categorized in terms of cost and benefit since they both represent the division of a cost by a benefit or vice-versa.

RC10, P14, L14: I think that the verb is “to take up”, and the noun “the uptake”.

RC11, P14, L17: The symbol “V_H2O” is the same as in Eq. (15) while its definition differs, which might be confusing for the reader.

RC12, P15, L24-25: I searched for references to the index “water yield” in the articles of Javaux et al. [2008] and Schneider et al. [2010], but only noticed that both display transpiration rate versus time, or mean water content, from which the cumulated transpiration rate at the onset of water stress could be estimated. Dividing this value by the total root length would provide water yield. Mentioning that these authors used the index “water yield” thus does not seem really accurate, and I believe that, under its current form, this index should be recognized as newly introduced (unless other literature would refer to it).

RC13, major comments about the interpretation of the index effort: Here I would like to discuss several points about the interpretation of the index effort.

Firstly, in the definition of effort, it is said that the value of effort (w~) is taken at the onset of water stress. Let’s consider two root systems with constant Tpot, one reaching early water stress with its (absolute) leaf xylem water potential increasing in a concave way, the other reaches water stress later with its leaf water potential increasing in a convex way and always lower than the other plant’s water potential. If effort is saved at the onset of water stress (a different time for the two plants), effort might very well be higher for the plant that permanently had the lower leaf water potential. With its current definition, and in the graphics displayed in the manuscript, low effort is thus not tantamount of low xylem water potential (in opposition to the statement at page 16, L7). I believe this is a big weakness in the definition of effort that might lead to misinterpretations in the results of this study. A simple way to dodge this weakness would be to save effort at the same time for all hydraulic architectures. Then a permanently lower (absolute) xylem water potential would always have for consequence a lower effort. This could also be applied to water yield by not excluding “stressed” transpiration rate and estimating water yield at a uniform time.

Secondly, only the root hydraulic architecture is supposed to change when studying the impact of root length and topology on effort. However, if the gravitational potential was accounted for, and the soil matric potential initially uniform with depth, the initial total soil water potential must have been initially different in simulations having different root depths. Particularly, soil water potential around the extremity of long roots was up to 500 hPa lower than around the extremity of short roots. The initial plant-sensed water potential was thus initially more negative for long roots. Such situation may have participated to the increase of effort for longer roots. An initial hydrostatic equilibrium would probably result in significantly different results (particularly no increased effort for longer roots) and be in better agreement with the assumption of neglecting vertical soil water flow (soil water flow is null at hydrostatic equilibrium while drainage occurs for a uniform soil matric potential). Actually, even using an initially constant soil matric potential and accounting for vertical water flow would progressively tend to equilibrate soil water potential in the root zone and below, thus draining water away from shorter roots, and would probably also result in the absence of increase of effort for longer roots.

These two points and RC7, make it possible that the chosen simplifications were responsible for the observation that effort increases with root length. In order to clarify the situation, I would advise the authors to repeat the simulations with (i) short and constant root segment lengths, (ii) an initial hydrostatic equilibrium, and (iii) to calculate effort at a uniform time for all hydraulic architectures, or to demonstrate that the increase of effort was a consequence of an increased (absolute) xylem water potential.

References:

Javaux, M., Schroder, T., Vanderborght, J., and Vereecken, H. (2008), Use of a three-dimensional detailed modeling approach for predicting root water uptake, Vadose Zone J., 7, 1079-1088.

Lobet, G., Pagès, L., and Draye, X. (2011), A novel image-analysis toolbox enabling quantitative analysis of root system architecture, Plant Physiol., 157, 29-39.

Lobet, G., and Draye, X. (2013), Novel scanning procedure enabling the vectorization of entire rhizotron-grown root systems, Plant Methods, 9, 10.1186/1746-4811-9-1.

Schneider, C. L., Attinger, S., Delfs, J. O., and Hildebrandt, A. (2010), Implementing small scale processes at the soil-plant interface - The role of root architectures for calculating root water uptake profiles, Hydrol. Earth Syst. Sc., 14, 279-289.

Referee Report: PDF

Hide

RR by Anonymous Referee #4 (10 Jul 2014)

Suggestions for revision or reasons for rejection

General comments
RC 1: Recent attention is given to the impact of root architectural and hydraulic properties on root water uptake from soil in different contexts, e.g. breeding for drought resistant crops (Vadez 2014), improved representation of root water uptake and transpiration in hydrological and climate models (Li et al. 2012) but data for parameterization of models are scarce. This manuscript raises some interesting questions within this topic. After reading the manuscript, I understand that the main new approach in this paper is as follows:
A new index “effort” is introduced, that determines, based on the porous pipe network assumption, how much work was required per unit root water uptake. It is shown that this index can be used to derive optimal root lengths for different distributions of root architectural and hydraulic parameters. Results can help answer the question whether a plant should invest in prolonging existing roots or in new branches based on the underlying hydraulic model and the assumption that this model is true.
However, I found the representation of the approach and results lengthy and in parts repetitive, and I had to struggle somewhat to filter out the really new issues. I suggest that the whole presentation of the paper should be more concise and compact. Repetitive parts should be removed; e.g. P5 L24-25 and P8 L21-23 / P15 L20 and P15 L22-23 and P21 L25-27 and P23 L7-8 / P8 L11 and P23 L25

RC 2: I find the title misleading: When I read the title I expected to learn about new ways to find parameters for root water uptake models, perhaps by an inverse modelling approach. Instead, I found that this paper evaluated different hypothetical distributions of root architectural and hydraulic parameters on the efficiency of root water uptake by a root and root system, respectively. The authors might consider a new title such as “Optimal root lengths in dependence of root architectural and hydraulic properties – Implications for root water uptake from soil”.

RC 3: I agree with anonymous referee #1 that the explanation of “efficiency” as it is used here could be improved. In particular, I also find the use of two indices that give similar results more confusing than helpful. When looking up the two given references for water yield (Javaux et al. 2008 and Zwieniecki et al. 2003) I could not find a definition of this term there. Of course both papers discuss water uptake over time but I don’t find this sufficient to state that water yield is a common index (while effort is a new one). In this context, I would also like to pick up on a comment of referee #1, namely the mentioning of water productivity (or water use efficiency). This is a well-known efficiency criterion, meaning efficiency of water uptake in terms of yield, and you could discuss the difference of your new (hydraulic?) efficiency criterion.

RC 4: I think it would also be interesting to compare this index “effort” with the “equivalent conductance of the root system” Krs as defined in Couvreur et al. (2012) as it directly depends on the root architectural and hydraulic properties. What would be the Krs value for the different parameter distributions that are found optimal with respect to effort?

RC 5: I did not understand why you would arbitrarily set root segments to be young or mature if you have the information about root segment age. You could still define two age classes such that you reach a desired length percentage in one and the other class, but the position of those would be realistic with respect to root segment age. Thus, the statement on P20 L10 that young roots are located anywhere within the root system should also be revised accordingly.

RC 6: I find the term “root strand” rather uncommon. I found only one other plant root system model-related publication that used this term, and I could not find it in the plant ontology database http://www.plantontology.org/. Do you mean “single unbranched root”?

RC 7: Another question regarding the use of terms: “redistribution of root water uptake” and “activation of mature roots” sound as if the plant would actively decide which parts of the root systems to choose for water uptake, as if it could be switched on or off. But is it not the case that the position where root water uptake occurs is a (passive) consequence of root architectural and hydraulic properties as well as the given scenario?

RC 8: The link between simple and complex model is not very clear. It is not the case that the results of the simple model are subsequently used in the more complex model, as I had expected after reading the section headings. Rather, the two examples are developed in parallel and it is not quite clear how and if to compare them. Would it not be possible to use the aRoot model with a single branched or unbranched root in order to have the same soil scenario for simple and complex branching structures?

Specific comments
P21 L22-31 is more a summary than a discussion.

P23 L23: By “metric”, you don’t mean a metric in the mathematical sense but an “index”?

P18 L16: I believe that the definition of moving uptake front (MUF) as in Levin et al. 2007 is defined in terms of the soil zone where RWU occurs. This is different than in your definition that defined it in terms of the position along the root where RWU occurs.

P27 L8: Q is a flow and not a flux as correctly stated on P 26 L21.

P28 L5: Is there a subindex “C” missing in the definition of EU?

New references
Li, L. et al., 2012. Improving the responses of the Australian community land surface model (CABLE) to seasonal drought. Journal of Geophysical Research: Biogeosciences, 117(G4): G04002.
Vadez, V., Root hydraulics: The forgotten side of roots in drought adaptation. Field Crops Research, in press, http://www.sciencedirect.com/science/article/pii/S0378429014000872#.

Hide

ED: Reconsider after major revisions (21 Jul 2014) by Nadia Ursino

AR by Marcel Bechmann on behalf of the Authors (03 Sep 2014) Author's response Manuscript

ED: Publish as is (05 Sep 2014) by Nadia Ursino

AR by Marcel Bechmann on behalf of the Authors (05 Sep 2014)