Classification of Hydrological Relevant Parameters by Soil Hydraulic Behaviour
Round 1
Reviewer 1 Report
The manuscript describes a method for soil classification based on the similar behavior of the soils with respect to their hydrological processes. The idea to classify soils on their hydrological behavior is however not new e.g. Groenendyk et al. [2015]. The authors just provide from my point of view an analysis of numerical data based on a larger sample size and higher temporal resolution. This clearly limits the novelty of the submitted manuscript of the authors. In addition important parts of the Material and Method section (2.3 and 2.4) are not well described and thus made it difficult to evaluate their results. The description of the Material and Methods section should enable other researcher to repeat and reproduce their work. I recommend the authors to work on the novelty of the manuscript e.g. in comparison to older publications. Perhaps it would make sense to include bulk density data, soil layering or showing apart from the numerical study also a real impact assessment for larger catchments.
Based on the submitted text I recommend at this stage a rejection of the manuscript. However I also want to encourage the authors to resubmit their manuscript after improving the novelty of their investigation.
General comments:
1) The novelty of the proposed investigation is missing. Groenendyk et al. [2015] already show a similar procedure for soil classification based on hydrological behavior.
2) Several very important information/aspects are missing in the Material and Methods section, which makes it in addition to point 1) difficult for me to fully review the submitted manuscript. Here are some of the main question which remained after reading the section 2.3 and 2.4:
a. How the authors generated the triplets of texture? To state a short algorithm randomly create the data is too sparse information.
b. Why using different simulation domains
c. No detailed description of the used numerical solver. The authors at least should describe the solver in detail and also describe why a solver is more adequate to use instead of a model like Hydrus-1D or similar models which also account for evaporation.
d. The time series are very short, so questionable if numerical data catch the entire hydrological behavior of the different soils.
e. The information on the lower boundary is totally missing. Which was the lower boundary and how was this chosen for each different set of parameters? The lower boundary plays especially for small vadose zone domains an important role on the hydrological behavior of water fluxes in the soil.
Detail comments:
Line 34: please clarify that smallest unit of model calculation is based on data availability from which scale? One local soil moisture observations for 1km², 1m², or 25px?
Line 52: In text citation….elaborated by [8]. Propose to write elaborated by Davis et al. 1927 [8]. Please change for the entire manuscript for in text citations.
Line 81: please clarify why water balance or change in water storage is a too general input for clustering and soil water flux not!
Line 85: please clarify why a higher number of numerical samples improve soil classification scheme?
Line 83-91: please state hypothesis or the objectives of the investigation. It is also quite unusual to give results in the introduction section e.g. line 91: “Results showed remarkably lower…..”
Line 99: please report somewhere in the M&M section also the lower boundary conditions used in the simulation.
Line 101: Wording does not fit .The expression of breakthrough curves are classically used in tracer/ solute transport studies. Please clarify and rewrite sentence.
Line 108: The equations of section 2.2 are all well known in the soil science or hydrological community
Line 110: Change W to With: below eq.1
Line 112: I recommend to changing capital theta to small theta for the entire manuscript.
Line 113: What is meant by simulated suction curve? I think the author mean simulated soil water retention curve/ characteristic
Line 117: Meaning of …… “with the 16 [16] equation”? Change to with the Richards equation [16]
Line 121: Unclear how the generation of texture triplets was conducted. Please describe in detail what the “short algorithm did”. This is necessary to make science reproducible.
Line 136: Ok the authors used a solver. However, the reader and review need to know more in detail what the solver did. Please describe it more in detail.
Line 140: Not sure what the authors did? Constant rainfall amounts of 2.5px per day. Or an impulse of one’s 2.5px water for the entire simulation (which would be very small in comparison to the water stored in the domain).
Line 141: the authors used also here a very short period 89 days and little rainfall amount. I recommend to at least using one year with a drying and wetting period. Additionally are such classifications dependent on the boundary conditions? Would you get the same classes for a very dry climate (200mm/a) or a very wet climate 1800mm/a. Perhaps it is worth to think in average and extreme conditions to cover soil classification based on hydrological behavior more globally.
Line 143: why different domains sizes?
Line 144: why choosing –L cm for initial conditions? Instead of using pressure heads at field capacity? This would connect the initial conditions on the different soil texture and make it more comparable.
Line 146: The description of the lower boundary is totally missing. This is with respect to the simulation of water fluxes in the domain an important setting.
Line 164: Please explain why the model efficiency of the single time series of water fluxes was compared to the mean water flux of all other time series within one class?
Line 168: What is E? Perhaps use NSE, which is a common abbreviation of the Mash-Sutcliff-Efficiency.
Line 218: Cahnge units of the Mualem van Genuchten parameter to θr and θs (cm³/cm³), α (1/cm). Check entire manuscript.
Dear authors: I didn´t furhter review the result and Discussion, because the section 2.3 and 2.4 are formulated too vaguely and novelty in the approach was missing
Groenendyk, D. G., T. P. A. Ferré, K. R. Thorp, and A. K. Rice (2015), Hydrologic-Process-Based Soil Texture Classifications for Improved Visualization of Landscape Function, PLOS ONE, 10(6), e0131299.
Author Response
Please see the attached file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Please look at attached review comments.
Comments for author File: Comments.pdf
Author Response
Please see the attached file.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper introduces a new classification of hydrological parameters based on soil hydraulic properties, which may be applied in simulations of soil water flux. The submission addresses one of key issues in hydrological and hydrogeological modelling. Consequently, it may be of wide interest for researchers dealing with the problem of possibly the most accurate parametrization in simulations of soil water movement. As such the paper deserves publication in the journal. In my opinion methodology applied by the Authors is correct and the obtained results prove relatively high reliability of the proposed classification. However, there exist some shortcomings (mostly technical), which require improvements prior to its publication. They are as follows:
1. Notations, as for example “[-]”, “[LT”, “[T]”, used when explaining equations are quite confusing. As I understand, they refer to units of the respective indicators applied in the equations. It is suggested to use units used in the International System of Units (SI) instead.
2. Figure 3: it is unclear which of the 15 coloured patterns corresponds to which class. I think that the respective patterns should be linked with the van Genuchten parameters shown in Table 1. So, it will be reasonable to add in Figure 3 a legend with a coloured pattern assigned to a corresponding number from Table 1.
3. The English language requires some improvements, preferably by an English native-speaker, who is additionally familiar with modelling in hydrology and hydrogeology.
Generally, I recommend the paper for publication after minor amendments mentioned above.
Author Response
Please see the attached file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Dear Authors,
I just revised manucsript and could not find a substantial improvement. There is from my point of view still the lack of novelty, despite the authors tried to improve the M&M section. The authors stated that the used in comparison to Groenendyk et al. 2015 the variable water fluxes and not the variable “change in soil water storage” and a longer real measured precipitation time series (which is by the way not a real atmospheric boundary, because evaporation is missing). However, why would the authors expect a large difference between results from Groenendyk et al. 2015 and their results?
It think the authors should think about including something new/novelty e.g. a real atmospheric boundary with evaporation (this is just neglected in their approach and depends also on the soil properties), using a deeper located bottom boundary as constant head impact water flux dynamics across the boundary (not suing evaporation will neglect capillary rise, which also depends strongly on soil properties), include bulk density data, a set- up with layered soils (more complex set-up and compare it with the simple) and so on
To conclude the topic of the study is of interest for hydrologist, soil scientist or other disciplines, but it is in my point of view still immature and needs more novelty in their set-up.
Best regards
Author Response
See the attached file.
Author Response File: Author Response.pdf
