Technical note: Saturated hydraulic conductivity and textural heterogeneity of soils

García-Gutiérrez, Carlos; Pachepsky, Yakov; Martín, Miguel Ángel

doi:https://doi.org/10.5194/hess-22-3923-2018

Articles | Volume 22, issue 7

https://doi.org/10.5194/hess-22-3923-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-22-3923-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 22, issue 7

Technical note

|

20 Jul 2018

Technical note |

| 20 Jul 2018

Technical note: Saturated hydraulic conductivity and textural heterogeneity of soils

Carlos García-Gutiérrez, Yakov Pachepsky, and Miguel Ángel Martín

Download

Final revised paper (published on 20 Jul 2018)
Preprint (discussion started on 05 Dec 2017)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'In attached and below', Robson André Armindo, 13 Jan 2018
- AC1: 'reply to referee #1', Carlos García-Gutiérrez, 17 Jan 2018
RC2: 'review1', John Koestel, 19 Jan 2018
- AC2: 'Response to the second referee comment.', Carlos García-Gutiérrez, 26 Jan 2018
  - AC3: 'Additonal response to reviewer #1 and #2', Carlos García-Gutiérrez, 01 Feb 2018

Peer-review completion

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

ED: Reconsider after major revisions (further review by editor and referees) (08 Feb 2018) by Roberto Greco

AR by Carlos García-Gutiérrez on behalf of the Authors (29 Mar 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (03 Apr 2018) by Roberto Greco

RR by Anonymous Referee #2 (04 May 2018)

Suggestions for revision or reasons for rejection

I am thanking the authors of hess-2017-706for their revisions, which have improved the manuscript. Unfortunatley, there are still plenty of critical points that require to be addressed before the manuscript may be accepted.

First of all, it strikes me that the excellent regression results are due to two reasons: a) the vast majority of soil samples in your data base are sand and sandy soils. Sand soils are known to lack stable textural superstructures that are formed in the presence of sufficient clay contents. It is likewise well known that Ks is therefore very well predicted from soil texture data in such soils. The second reason that helps increasing your R2 (and decreasing the RMSE) is that you are binning the data into 10 classes. This removes a lot of intra-class variability. You therefore must not mention the R2 value in the abstract as it is artificially increased. Likewise, the abstract must better reflect the fact that the regression results are very good for the sandy soils but clearly worse for the finer textured soils (as it is very much expected from previous studies). By the way: I do not understand why it is necessary to apply the binning. It needs to be explained in the manuscript.

I still do not think that including Ks is necessary. It only complicates understanding what you have done. Log Ks is enough. If you do not agree please explain in the manuscript why both are needed. The authors moreover need to make clear whether they are using log_{10} or ln [i.e. log_{e}). the log-notations are furthermore inconsistent throughtout the manuscript.

The authors state that it is the aim of this study to test whether the combination of different texturial triplets and the information entropy yields "a better description of K_sat" (p2L16). It is not clear to me what is meant with "description". I am interpreting it as "prediction" from texturial data. As also stated by the AE, the authors are statistically predicting Ksat from texturial data. They are providing very little physical reasoning/theory for why their approach should work better. In this, they are basically creating pedotransfer functions. In fact, the authors are apparently failing to take soil structure into account, i.e. stable superstructures built from individual texturial particles that are held together by clay sized particles. Such structures are often biopores e.g. dug by earthworms or old root channels. In the "final comments" At P8L2 the authors are writing that "soil aggregation, root exudates, soil aggregation, etc." are other important factors that determine Ks. I completely agree. However, this must already be discussed in the introduction. It will help the authors to explain why the texturial data is performing so much worse for the finer textured soils.

I have for now assumed that the database used by the authors contains undisturbed samples. It is not stated in the manuscript, though. It needs to be stated there because it is a very important information for understanding the results (see comment above).

Another major concern is the motivation of using the information entropy. It is simply not provided in the manuscript. It is unclear to me why it should be better then using a texturial triplet directly for a regression. An explanation should be given and a respective reference included. The authors aim at testing whether it would yield "a better description of Ksat". They are then however only testing the effect of different texturial triplets. If also the information entropy should be tested for usefulness for predicting Ksat, a comparison to linear regressions without using the information entropy needs to be included. I am urging the authors to do this. It would then be interesting to discuss why the IE is working better. Maybe because it tends to linarize the relationship to log Ksat?

I propose another round of major revisions.

More specific comments:

P1L6-13: I agree that the particle size distribution has a large impact on Ksat. Please explain better that it is the soil pore network architecture that determines Ksat (not the texture itself) and that stable structural pores (i.e. biopores, cracks, etc.) are only existing if the clay content is large enough to support them. Please quote some respective references, e.g. Horn, R., H. Taubner, M. Wuttke and T. Baumgartl. 1994. Soil physical properties related to soil structure. Soil Tillage Res. 30: 187-216. Explain that the structural pores are often more important for Ksat because they are larger than the inter-particle pores (q ~ r^2) and mostly well connected. Furthermore explain what advantages the use of the IE brings and what the findings of Boadu (2000) were. Was IE found to be an improvement?

P3L17-19: justify the use of the binning approach. The reader should not get the impression that it was only used to make the results look better.

P6L21: not just "possibly aggregating". In undisturbed soil the fines are as good as always aggragated. Exception are highly sodic soils.

P6L23-34: the scale breaking discussion is interesting but needs to be expanded. I still do not find it convincing that what you observe is caused by the scale breaking. Figure 6 shows that the relationship between IE-3-1-3 and log Ksat is non-linear (not only "could be" linear). If you would use a quadratic regression, you would again get very nice fits.

P7L14-23: move to the materials and methods section

P8L17-20: I do not understand what the authors are trying to state here. First of all, such a passage belongs to the end of the introduction, where the aim of the study is defined. Secondly, PTFs may be linear or non-linear, but they are not created for interpretation. Moreover, this manuscript does does only minimally involve physical relationships.

Hide

RR by Robson André Armindo (07 May 2018)

Suggestions for revision or reasons for rejection

The manuscript was improved in several aspects, mainly in written part giving more details to the readers. I am glad that some of my previous points were helpful to the authors as I may see over this new version. I see also that new graph (Fig 6) and analysis are in this version. By the authors explaining, I understand now that this work does not intend to develop PTFs to estimate Ksat but to present a physical explaining of the relationship between Ksat and PSD. If so, this information should be clearer in the text.

Plus, I still have two main comments in order to help authors to get a better text:

- It is pointed out as objective that this work yielded a better description of Ksat. However, I cannot find in conclusion item which was this description of Ksat. I may see that PSD, by means of sand, silt and clay sizes, correlated with Ksat values. In my point of view this is a statistical correlation and not a physical characterization. So, if this work has no provided a physical characterization of Ksat for explaining it, the objective could be rearranged to be presented in another way. In conclusion item is written that PSD correlate with Ksat. Was the aim of this work to verify a possible correlation between both variables or to physically describe Ksat in function of PSD?
- Fig. 6 shows in title ln Ksat and in graph log Ksat. Please, be careful with that. Authors should fix it over all text.

Minor comments
- Authors are still using the symbol “.” over the text to represent values and I think in ESL “,” would better describe it. For example, at line 21 where is “19.000” I think it should be “19,000”.
- Again, I wonder to know if the name “heatmap” is the better name to those maps that do not show values of heat or temperature. Is it because of entropy information? I had no any reply about this previous mentioned query.
- The quality of Figures 1, 2, 4, 5 and 6 can be improved.
- Variables Ksat and RMSE presented in Table 4 and 5 are missing units.
- Table 5. log10Ksat or lnKsat?

Hide

ED: Publish subject to revisions (further review by editor and referees) (08 May 2018) by Roberto Greco

AR by Carlos García-Gutiérrez on behalf of the Authors (14 May 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (18 May 2018) by Roberto Greco

RR by Anonymous Referee #2 (25 Jun 2018)

ED: Publish as is (28 Jun 2018) by Roberto Greco

Short summary

Saturated hydraulic conductivity (K_sat) is an important soil parameter that highly depends on soil's particle size distribution (PSD). The nature of this dependency is explored in this work in two ways, (1) by using the information entropy as a heterogeneity parameter of the PSD and (2) by using descriptions of PSD in forms of textural triplets, different than the usual description in terms of the triplet of sand, silt, and clay contents.