the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Apr 2022
On soil bulk density and its influence to soil moisture estimation with cosmic-ray neutrons
Abstract. Cosmic-ray neutron sensing (CRNS) is a non-invasive technique that is used to quantify soil moisture in a representative footprint of 10–20 ha and 15–80 cm depth. In a stationary mode, CRNS is widely employed to monitor the moisture dynamics on agricultural land, which may undergo active changes of the soil compartments, e.g. due to plowing. On mobile platforms, CRNS measurements aim at mapping the spatial soil moisture distribution across various types of soil and land use. To date, the potential effect of variable soil bulk density on the neutron measurements has not been investigated in detail. In fact, most sensors are calibrated only once on site-specific soil properties. Therefore we hypothesize that unaccounted spatiotemporal changes of soil bulk density may have impact on the quality of CRNS soil moisture products.
In this study, we quantify the effect of the soil density on the neutron response by neutron transport simulations and a dedicated lab experiment. The results indicate a significant dependency of neutrons on soil bulk density, which also depends on the soil moisture state. For the hypothetical cases with constant ratio between soil density and water content, the neutron intensity remains unaffected.
Correction functions are proposed to improve the performance of two widely used approaches to convert neutrons to soil moisture (Desilets et al. 2010, and Köhli et al. 2021). The latter approach together with the proposed correction can be recommended for practical use, as it accurately represented the simulated neutron response to soil moisture and soil bulk density with a constant – and potentially universal – calibration parameter.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(474 KB)
-
Supplement
(9 KB)
-
This preprint has been withdrawn.
- Preprint
(474 KB) - Metadata XML
-
Supplement
(9 KB) - BibTeX
- EndNote
Interactive discussion
Status: closed
-
RC1: 'Comment on hess-2022-123', Anonymous Referee #1, 10 May 2022
GENERAL COMMENTS
This paper comprehensively addressed the effects of soil bulk density on soil moisture estimation using cosmic-ray neutron sensing technique. The authors did a series of neutron simulations on a wide spectrum of soil bulk density and soil moisture. The simulation results were verified by conducting lab experiments of 4 scenarios sequentially. The effects of the bulk density were manifested by applying two widely used conversion equations on the simulation results. The authors also provide a correction approach for the CRNS community.
Overall, I think the authors did an acceptable job on simulation and lab experiment. However, the current version of the manuscript is not coherent or cohesive. There is plenty room for improvement in results presentation, content organization, and writing. Detailed suggestions and corrections are listed below.
SPECIFIC COMMENTS
Relationship to lattice water and organic matter
I am curious about if any effects of lattice water that has already been considered in the lattice water correction are considered again in the bulk density correction. The case with organic matter is similar. As mentioned in Line 267-269, mineral soils without any organic parts were considered. However, the low bulk density of soil is usually the consequence of high content of organic matter. In practice, if I apply both the organic matter correction and bulk density correction to my neutron counts, do I over-correct it?
Objectives and conclusions
The last two paragraphs of the introduction section need to be largely modified or rewritten. The hypothesis is only mentioned in this section but was not explained in detail how it is rejected/accepted and according to what criteria. To make the conclusion address the introduction, the objectives should be somewhat like 1) investigate the influence of bulk density in CRNS soil moisture measurement, 2) quantify the influence by examining the sensitivity of neutron counts through a large spectrum of soil bulk density, 3) assess the impact of bulk density on the two major conversion functions, and 4) develop a novel correction method/parameter for practical use.
Presentation of the results
The study provides substantial amount of simulation results. Many figures are clearly plotted and very informative. However, I feel a little awkward reading some of the figures. Based on the simulation and experiment design, I expected more figures like Fig. 5 with soil moisture as x-axis, and porosity contour lines, instead of the opposite, like Fig. 3. The presentation of Fig. 5 is similar to Fig. 6 in Zreda (2012) which is probably more common and easier to read in published literature.
The lab experiment is of great value to this study, which could potentially provide a great validation dataset for the simulation. I wonder if there is any experiment on varying soil moisture conducted on this setup. It would be great if there were more points on Fig. 3 or a separate figure of Nexp vs. Nsimu.
For Fig. 3 specifically, only 2 of the 4 experiment results were plotted in the figure. They were not colored with the soil moisture color scheme, which can barely inform the readers with the performance of the simulation/experiment. I suggest plotting a zoom-in panel of the four points with four ideal contour lines of the soil moisture content equal to the four measurements. The four points should also be colored with same color scheme.
Notation
The notations in this paper are generally clear but some selections are not very common. For dry bulk density, I think ρb is more common and ρs is often used as the density of solids instead of bulk density, which may cause some confusion.
For Section 3.1, “soil-water ratio” may be better named as “solid-water ratio”. You may assign this ratio with a symbol since it is also used in Table 1 and Section 3.3.
The correction factors
Some of the correction factors depend on soil moisture, which doesn’t really “go in line with the traditional correction factors for atmospheric changes…” (as mentioned in Line 337), since I think the correction factors should be independent of soil moisture. Line 374-375 mentioned a TDR sensor. Does it mean that a point-scale measurement or some rough estimate can also work here?
TECHNICAL CORRECTIONS
(not a complete list)
Line 28: in the soil -> In the soil
Line 98: Add definition of N0
Line 150: what is “DIN 18 125 T1 and T2”?
Line 176: Describe the y-axis. Is the “simulated neutron counts” with a. u. (arbitrary unit?) the same with the “neutron counts (normalized)”?
Line 208: section 2.2 -> Section 2.2 (many similar ones throughout the manuscript)
Line 209: Add “ in Koe. Function” before “(Eq. 8)”
Line 214: Extra space before “Eqs. 5”
Line 227-228: delete “above” and “less then”
Line 249: Fig. 3) -> Fig. 3a)
Line 254: (Fig.3) -> Fig. 3b)
Line 260-261: Any reference for the classification of bulk density? If the medium dense soil is between 1.6 and 1.8, this classification is not complete. For example, 1.5 g/cm3 doesn’t belong to any category.
Line 296: (6) -> Fig. 6
Line 340: section 2.2 -> Section 2.1
Line 356: based on; is only depend on -> only depends on
Line 358: the logic chain is broken between the two sentences
Line 370: extra “an”
Line 400: delete “small”
Figure 1: Instead of randomly giving 2 examples of the 7 simulations, describe the 7 simulations with the aid of Figure 1.
Table 1: Description of this table is not complete in the text. The column "differential" has different meanings across rows since the last row is in percentage. It is also mentioned as “deviation” in the text.
For Section 3.3, does it aim to verify/validate some of the simulation results? If so, consider renaming this section. The “previous section” (Line 206) is Section 3.2 or 3.1? If 3.1, consider switching/merging Section 3.2 and 3.3.
Citation: https://doi.org/10.5194/hess-2022-123-RC1 -
AC1: 'Reply on RC1', Martin Schrön, 24 Jul 2022
Dear Reviewer 1,
thank you very much for your positive review. We will briefly respond to your concerns in the interactive discussion. (RC=reviewer comment, AR=author response)
# Main comments
> RC1.0: "Overall, I think the authors did an acceptable job on simulation and lab experiment. However, the current version of the manuscript is not coherent or cohesive. There is plenty room for improvement in results presentation, content organization, and writing. Detailed suggestions and corrections are listed below."
AR: Thank you for your critical review and the suggestions to improve the manuscript. We will consider to restructure the text and to improve the writing throughout the revised manuscript.
##Specific comments
> RC1.1: "Relationship to lattice water and organic matter: I am curious about if any effects of lattice water that has already been considered in the lattice water correction are considered again in the bulk density correction. The case with organic matter is similar. As mentioned in Line 267-269, mineral soils without any organic parts were considered. However, the low bulk density of soil is usually the consequence of high content of organic matter. In practice, if I apply both the organic matter correction and bulk density correction to my neutron counts, do I over-correct it?"
AR: There is no risk of over-correction when correction for water-equivalent and bulk density are performed separately. As we have described in L43 and L269, lattice water and organic matter translate to an equivalent of water density, which is taken into account by the variable θ.
In general, epithermal neutrons are mainly sensitive to the density ratio between water and solid soil (as shown in section 2.1). They are rather insensitive to other properties of soils and molecular structure (see, e.g., Zreda et al. 2008, 2012 or Köhli et al. 2015). Hence, only those two components (soil density and hydrogen content) are to be considered when processing neutron data. In literature, soil density has not been considered yet (other than converting from gravimetric to volumetric), which is the motivation for the present study. Soil's hydrogen content has to be corrected for the hydrogen in lattice water as well as in organic material, in order to end up with a quantity representing mobile water/soil moisture only (see line 43). In the state of the art, this is done by simply substracting lattice water and organic water equivalent from the total water content measured by CRNS. The reviewer is right that one should correct for these variables, but this will only correct for hydrogen-related effects and not for any bulk density effects. Hence, these corrections can be treated independently.
On the other hand, we fully agree that low bulk densities are often correlated with high organic matter content. The value pairs of high porosity and low (effective) water content are therefore not very common in typical landscapes. In L269 we have already indicated this concept (Soils with bd = 1 g/cm³ have low abundance and commonly include a high organic carbon content which introduces an additional effect to θ."). We will make more clear that the soil moisture values used in the equations and in the figures comprise the total hydrogen content, i.e., mobile soil water + lattice water + organic water equivalent.
> RC1.2: "Objectives and conclusions: The last two paragraphs of the introduction section need to be largely modified or rewritten. The hypothesis is only mentioned in this section but was not explained in detail how it is rejected/accepted and according to what criteria. To make the conclusion address the introduction, the objectives should be somewhat like 1) investigate the influence of bulk density in CRNS soil moisture measurement, 2) quantify the influence by examining the sensitivity of neutron counts through a large spectrum of soil bulk density, 3) assess the impact of bulk density on the two major conversion functions, and 4) develop a novel correction method/parameter for practical use."
AR: Thank you for suggesting a clearer structure for the outline of the manuscript, we will consider adapting the last paragraphs of the introduction section accordingly. The reason why we initially decided to start with the negative hypothesis -- that neutrons are invariant against change of soil bulk density -- was that this has been the hitherto common assumption in CRNS literature. However, we understand that a more straight-forward description of the hypotheses would be more clear.
> RC1.3: "Presentation of the results: The study provides substantial amount of simulation results. Many figures are clearly plotted and very informative. However, I feel a little awkward reading some of the figures. Based on the simulation and experiment design, I expected more figures like Fig. 5 with soil moisture as x-axis, and porosity contour lines, instead of the opposite, like Fig. 3. The presentation of Fig. 5 is similar to Fig. 6 in Zreda (2012) which is probably more common and easier to read in published literature."
AR: Thanks for sharing your impression regarding the presented figures. We agree that neutrons-over-moisture figures are more common in CRNS literature, which is why we decided to plot Fig. 5 this way to illustrate the effect for a typical application. But the main body of the study is the investigation of the measurement variable (neutrons) in responds to the changing variable (soil bulk density). As it is common sense to plot the independent variable on the x-axis, and the dependent variable on the y-axis, we believe that neutron-over-density plots most clearly show the dependency of neutrons on bulk density, while the additional dependency on water content is of secondary importance.
> RC1.4: "The lab experiment is of great value to this study, which could potentially provide a great validation dataset for the simulation. I wonder if there is any experiment on varying soil moisture conducted on this setup. It would be great if there were more points on Fig. 3 or a separate figure of Nexp vs. Nsimu. For Fig. 3 specifically, only 2 of the 4 experiment results were plotted in the figure. They were not colored with the soil moisture color scheme, which can barely inform the readers with the performance of the simulation/experiment. I suggest plotting a zoom-in panel of the four points with four ideal contour lines of the soil moisture content equal to the four measurements. The four points should also be colored with same color scheme.
AR: Thank you for the suggestions. We will improve this figure visually and we will consider adding more data points if applicable.
> RC1.5: "Notation: The notations in this paper are generally clear but some selections are not very common. For dry bulk density, I think ρb is more common and ρs is often used as the density of solids instead of bulk density, which may cause some confusion. For Section 3.1, “soil-water ratio” may be better named as “solid-water ratio”. You may assign this ratio with a symbol since it is also used in Table 1 and Section 3.3."
AR: Thanks for the suggestions. From the physics point of view (section 2.1), dry soil bulk density is equivalent to the solid part of the soil, so we used abbreviations s (solid), w (water), and a (air) for the three main components, and we tried to be consistent with this notation in the remaining manuscript. Using ρb for soil bulk density would require a redefinition of ρs = ρb, which would be unnecessary and redundant. But we understand that it would be helpful to use the notations more common in soil science. Hence, we will reconsider the notation of dry soil bulk density in the revised manuscript.
> RC1.6: "The correction factors: Some of the correction factors depend on soil moisture, which doesn’t really “go in line with the traditional correction factors for atmospheric changes…” (as mentioned in Line 337), since I think the correction factors should be independent of soil moisture. Line 374-375 mentioned a TDR sensor. Does it mean that a point-scale measurement or some rough estimate can also work here?"
AR: We fully agree that the moisture-dependent correction is unconventional and may lead to a circular problem. The approach for correcting the road effect showed similar characteristics (Schrön et al. 2018). We tried very hard to find a solution that is independent of soil moisture, but unfortunately we cannot "work against the rules of nature". While moisture-dependent correction leads to the best performances, the good news is that moisture-free approaches also showed acceptable performance -- if the rough moisture regime is known (see the first two rows in Tables 2 and A1). So, yes, a very rough estimation of soil moisture is sufficient to achieve a decent correction performance. We will stress out this aspect more clearly in the revision.
# Technical corrections
AR: Thank you very much for your time to hint at technical issues and for making adequate suggestions. We will consider them thoroughly in the revised manuscript.
Citation: https://doi.org/10.5194/hess-2022-123-AC1
-
AC1: 'Reply on RC1', Martin Schrön, 24 Jul 2022
-
RC2: 'Comment on hess-2022-123', Anonymous Referee #2, 13 Jun 2022
Overall Comments
This paper presents an important topic for the calibration of CRNS sensors – whether or not there is a dependence on the soil bulk density. The issue has been somewhat neglected in the application of CRNS for soil moisture sensing, therefore this is an important study, especially for monitoring sites with tillage, or mobile CRNS applications, surveying across different soil types and bulk densities. The study is in the form of neutron transport simulations (using the URANOS model) and a dedicated lab experiment.
The presentation is reasonably good, with some additional details required. The English language is generally quite good, but should be improved. However, as it appears to me, there seems to be some major inconsistency between the theory and the presentation of results, and at times the text contradicts the plotted results. Attempting to untangle this is a major distraction from the message of the paper. Moreover, as presented currently I believe that the results are most likely incorrect, but will be corrected easily.
Detailed Points:
- The Discussion and Results present the increase of neutron counts with increasing bulk density (decreasing porosity) – whereas the theory (Equation 3) suggests that solid matter will scatter more efficiently than the much lower density air in the pore spaces, which implies to me that the neutron count should decrease with increasing bulk density (BD). Perhaps I am wrong, and misunderstand something here? But the text is also contradictory in places e.g. P.7, L.176-77:
“The highest neutron intensity can be achieved for very dry and dense soils, while it decreases with increasing soil bulk density (or decreasing porosity).”
If Fig.2 is plotted correctly, then this should at least be corrected to “….very dry and low density [loose] soils…”
However, I have doubts whether the plots are correct? For Fig.2 & Fig.3 (and subsequent plots with BD/porosity x-axis) – I believe that all the x-axes labels are reversed? Making this assumption, the results would make physical sense, otherwise, as above the results do not match (or are opposite to) the theory presented, as I understand it.
The later text does match the plots, so I am really left wondering if all the discussion and presentation of results has to be revised, as well as the plots… or else explain what I am missing and why we should expect higher neutron counts with higher BD?
- The paper is written in the future tense (e.g. P.2, L.54-55), this not the conventional way to report scientific experiments.
- Take more care with the use of tenses, and prepositions e.g. Title change “influence to soil moisture” to “influence on soil moisture”. English should be generally improved and errors removed.
- Section 2.4 Experimental Concept: Where are the tanks located? - how did you control the surrounding environment? Could there have been other non-constant hydrogen pools in the surrounding 100 to 200m?
- 5, L.139-L.140 specify the CRNS detector type (make & model). Why only a 60 minute counting period? Why not longer to improve the count statistics?
- Some figure captions (Fig.2 etc.) have the wrong units for BD – should be g/cm^3 not g/m^3.
- 9, L.191 were the neutron counts first corrected in any way? For example, for changes in atmospheric pressure, air humidity, incoming neutron intensity? Otherwise, neutron counts made at different times cannot be directly compared!
- Table 1. Are the neutron counts assigned to the correct porosities? Whilst data here matches Fig.3, it does not agree with the theory (Eq. 3)
- 11, L.240-242 Is this the right way round? Here both BD and SM vary - so which has the more dominant effect on penetration depth?
- 13, L.293 understimate (?) --> increased porosity should increase counts! Leading to underestimate of SM??.... and Fig.5 Why don't counts increase with porosity? I don't think this is correct?
- Summary & Conclusions: (2.) “On average, neutron count rates decrease by −1% for every +10% increase in porosity”. This appears contrary to Eq.3 ? I would expect an increase in count rate. As per Point 1 above, much of this text needs revising if indeed the plot x-axes are reversed.
Citation: https://doi.org/10.5194/hess-2022-123-RC2 -
AC2: 'Reply on RC2', Martin Schrön, 24 Jul 2022
Dear Reviewer 2,
thank you very much for your critical and positive review. We will briefly respond to your concerns in the interactive discussion. (RC=reviewer comment, AR=author response)
# Overall comments
> RC2.1: "The presentation is reasonably good, with some additional details required. The English language is generally quite good, but should be improved. However, as it appears to me, there seems to be some major inconsistency between the theory and the presentation of results, and at times the text contradicts the plotted results. Attempting to untangle this is a major distraction from the message of the paper. Moreover, as presented currently I believe that the results are most likely incorrect, but will be corrected easily."
AR: Thank you for sharing your impression. We believe that the reviewer misinterpreted the slope of the curve of equation 3, which led to an apparent inconsistency with the rest of the text and results. In the detailed response below we will try to solve this misconception, proving that the theory is indeed consistent with the simulations.
Large parts of this review refer to apparent inconsistencies of the figures, texts, and conclusions with equation 3. These comments are seemingly all based on an initial misconception, which can be quickly solved (see next section). However, we admit that a typo in line 177 ("decreasing" instead of "increasing") may have led to this substantial confusion. We apologise for this mistake will be fix the sentence in the revision.
> RC2.1: "The Discussion and Results present the increase of neutron counts with increasing bulk density (decreasing porosity) – whereas the theory (Equation 3) suggests that solid matter will scatter more efficiently than the much lower density air in the pore spaces, which implies to me that the neutron count should decrease with increasing bulk density (BD). Perhaps I am wrong, and misunderstand something here?"
AR: Your are referring to equation 3 (or the simplified equation 4), where N is proportional to a combination of cross sections, densities, and energy loss quantifiers. Here, the slope of the curve (N over s) highly depends on the weighting factors ξs and ξw in the denominator. It is 1 for ξs = ξw, <1 for ξs > ξw, and >1 for ξs < ξw. According to Köhli et al. 2021, the variable xi represents the energy loss per collision, and can be estimated with ξ=2/(A+1), where A is the atomic mass number. For hydrogen, ξ(A=1)=1, and for Silicon, ξ(A=28)=0.07. Hence, ξs << ξw.
For a quick illustration of how this influences N, let a=s*Σs and b=w*Σw, then N ~ (a+b)/(a*ξs + b*ξw). With ξs << ξw this reduces roughly to N~(a+b)/b = 1+a/b, i.e., N ~ s. Using rough values for all the variables, the following link shows this relationship plotted: https://tinyurl.com/neutronsoliddensityplot
Since the theory indeed confirms our simulation results, we can be assured that the plots and conclusions are correct. However, we realize that better understanding of the theoretical part is critical for the readers, so we will consider improving the explanations and guidance on interpreting the theoretical concepts.
> RC2.2: "But the text is also contradictory in places e.g. P.7, L.176-77:
“The highest neutron intensity can be achieved for very dry and dense soils, while it decreases with increasing soil bulk density (or decreasing porosity)."AR: You are right that this sentence contradicts itself. We meant, that neutron intensity *increases* with increasing soil bulk density. Sorry for the confusion, it will be changed in the revision.
> RC2.3: "However, I have doubts whether the plots are correct? For Fig.2 & Fig.3 (and subsequent plots with BD/porosity x-axis) – I believe that all the x-axes labels are reversed? Making this assumption, the results would make physical sense, otherwise, as above the results do not match (or are opposite to) the theory presented, as I understand it.
The later text does match the plots, so I am really left wondering if all the discussion and presentation of results has to be revised, as well as the plots… or else explain what I am missing and why we should expect higher neutron counts with higher BD?AR: The plots are correct and do confirm the theory (see above). The text will be double-checked for potentially confusing formulations.
> RC2.4: "The paper is written in the future tense (e.g. P.2, L.54-55), this not the conventional way to report scientific experiments.
Take more care with the use of tenses, and prepositions e.g. Title change “influence to soil moisture” to “influence on soil moisture”. English should be generally improved and errors removed."AR: Thank you, we will improve the writing upon revision of the manuscript.
> RC2.5: "Section 2.4 Experimental Concept: Where are the tanks located? - how did you control the surrounding environment? Could there have been other non-constant hydrogen pools in the surrounding 100 to 200m?
AR: The tanks were located inside a large, air-conditioned hall, build and surrounded by concrete. This way we expect no influence external weather conditions. Between 5 and 50 meters, scientists and cars have been present but with only minor rearrangement during the experiment period. This might be a valid source of systematic uncertainty, but it is constant and very hard to quantify. We will elaborate on the measurement setup in more detail.
> RC2.6: "5, L.139-L.140 specify the CRNS detector type (make & model). Why only a 60 minute counting period? Why not longer to improve the count statistics?
AR: The detector has been specified as the one used for roving in Schrön et al. 2018, with roughly 10x more count rate than CRS1000 detectors. The measurement period is short because we tryed to minimize the risk of water loss due to evaporation during the day. Our main goal was to keep the soil water content constant between the two experiments (dense and loose), so we removed the soil and repacked it within one day, leaving only about 60 minutes time for each of the two measurements. We will better describe this approach in the text.
> RC2.7: "L.191 were the neutron counts first corrected in any way? For example, for changes in atmospheric pressure, air humidity, incoming neutron intensity? Otherwise, neutron counts made at different times cannot be directly compared!"
AR: Yes, we corrected the neutron counts according to standard procedures. We will add this information to the revised manuscript.
> RC2.8: "Table 1. Are the neutron counts assigned to the correct porosities? Whilst data here matches Fig.3, it does not agree with the theory (Eq. 3)"
AR: We are glad to report that the data is consistent with equation 3, see the explanation above.
> RC2.9: "L.240-242 Is this the right way round? Here both BD and SM vary - so which has the more dominant effect on penetration depth?"
AR: We agree that this example is challenging to interpret. We will provide a more clear example that shows the different contributions of soil moisture and bulk density on the penetration depth. However, this is just a side note and based on already published literature (see D86 in Schrön et al. 2017).
> RC2.10: "13, L.293 understimate (?) --> increased porosity should increase counts! Leading to underestimate of SM??.... and Fig.5 Why don't counts increase with porosity? I don't think this is correct?
Summary & Conclusions: (2.) “On average, neutron count rates decrease by −1% for every +10% increase in porosity”. This appears contrary to Eq.3 ? I would expect an increase in count rate. As per Point 1 above, much of this text needs revising if indeed the plot x-axes are reversed."AR: Also here the text is consistent with the results and theory (see above).
Citation: https://doi.org/10.5194/hess-2022-123-AC2
-
EC1: 'Comment on hess-2022-123 - Start interacting', Nunzio Romano, 21 Jun 2022
While waiting for another reviewer's evaluation, I suggest you should start providing some preliminary responses to the comments received so far. This helps feed the discussion step of the journal.
Citation: https://doi.org/10.5194/hess-2022-123-EC1 -
RC3: 'Comment on hess-2022-123', Anonymous Referee #3, 26 Jun 2022
The manuscript reports a numerical and experimental study of the effect of soil bulk density on cosmic ray sensing (CRNS). Such an impact is important if the land surface is subject to farming practices and for applications of the CRNS method for mapping soil moisture across heterogeneous landscapes.
I believe that the impact of soil bulk density is an impact factor that needs to be considered at least for some applications when converting neutron counts into soil moisture and as such, the study is highly relevant and important and ultimately warrants publication. However, the manuscript is not ready for publication in its present form as far as the analysis, presentation and language are concerned. The authors need to make a thorough effort in improving the text and make it consistent both internally and in relation to the presented figures. During my reading of the manuscript, I noted language issues at several places but I will only mention some of them below.
Comments:
l. 50: Perhaps a bit unusual to start out with the hypothesis that that neutron intensity is unaffected by soil bulk density. The whole idea of the manuscript is to document that the opposite is the case, so it would be more intuitively to state that the objective of the work is to analyze and quantify the impact of soil bulk density on neutron counts and how to take this impact into consideration to obtain reliable estimates of soil moisture.
Section 2.4 Experimental concept: I believe that the reported experiments are very interesting in order to gain experimental evidence of the impact of soil bulk density. The experiments are indeed a comprehensive endeavor and I would like to know a bit more details on the setup perhaps by an accompanying figure, which also shows the position of the CRNS detector.
I anticipate that the packings were done with dry sands. How did you then add water afterwards to obtain uniform moisture contents in the packings?
Could there be environmental factors affecting the results?
l. 142: I suppose this should be for the loose experiment?
l. 155: In line 126 it is stated that the grain density in the simulations is 2.86 g/cm3 but now a value of 2.65 g/ cm3 is assumed? Any inconsistency here?
Figure 1: In my printout, water and air have the same signature. Please make the two phases mutually discernable.
l. 164: It is 4 units of soil and 1 unit of water giving a ratio of 4. Please be consistent how you refer to this ratio throughout the manuscript.
l. 165: equal 20 vol% porosity?
l. 177: This sentence is contradictory to Figure 2. Intuitively I would think the sentence is right.
l. 183: I am not sure I see a flat slope in Figure 3.
l. 186: Again, I would not call it a flat curve.
Figure 3b: In Figure 3a the counts become higher for lower soil moisture contents but the opposite is the case in Figure 3b, which does not make sense.
Section 3.3 Evidence in the sand box experiment: There are some conflicting statements in this section regarding counts. Do the counts increase significantly (l. 193) or are they the same (l. 205)? What is the learning from the experiments given that some the controlling parameters are not constant?
l. 284-300: Please revise these lines as well as the figures. There are conflicting figure numbers and porosity numbers and some of the referred numbers of counts and soil moisture I do not see.
l. 315: Impressive perhaps a bit exaggerated.
l. 324-325: This is an important statement which perhaps should be emphasized even more.
Citation: https://doi.org/10.5194/hess-2022-123-RC3 -
AC3: 'Reply on RC3', Martin Schrön, 24 Jul 2022
Dear Reviewer 3,
thank you very much for your critical and detailed review. We will briefly respond to your concerns in the interactive discussion. (RC=reviewer comment, AR=author response)
# Comments
> RC3.0: "I believe that the impact of soil bulk density is an impact factor that needs to be considered at least for some applications when converting neutron counts into soil moisture and as such, the study is highly relevant and important and ultimately warrants publication. However, the manuscript is not ready for publication in its present form as far as the analysis, presentation and language are concerned. The authors need to make a thorough effort in improving the text and make it consistent both internally and in relation to the presented figures. During my reading of the manuscript, I noted language issues at several places but I will only mention some of them below."
AR: Thanks for sharing your impression on the manuscript. We thank the reviewer for pointing out some language and writing issues, which will be solved during the revision. However, we have not been able to identify issues with the analysis based on this review.
> RC3.1: "l. 50: Perhaps a bit unusual to start out with the hypothesis that that neutron intensity is unaffected by soil bulk density. The whole idea of the manuscript is to document that the opposite is the case, so it would be more intuitively to state that the objective of the work is to analyze and quantify the impact of soil bulk density on neutron counts and how to take this impact into consideration to obtain reliable estimates of soil moisture.
AR: We agree and will reformulate the last paragraphs of the introduction according to the suggestions from Reviewer 1. The reason why we initially decided to start with the negative hypothesis (that neutrons are invariant against change of soil bulk density) was that this has been the hitherto common assumption in CRNS literature. However, we understand that a more straight-forward description of the hypotheses would be more comprehensive.
> RC3.2: "Section 2.4 Experimental concept: I believe that the reported experiments are very interesting in order to gain experimental evidence of the impact of soil bulk density. The experiments are indeed a comprehensive endeavor and I would like to know a bit more details on the setup perhaps by an accompanying figure, which also shows the position of the CRNS detector."
AR: Thanks for the suggestion, also the other reviewers asked for a more elaborate description of the experiment. Initially, we did not plan to elaborate more on the experiment, since the main focus of this manuscript is the theory (and simulations). But we agree that the lab experiment is a unique setup which deserves more detailed explanations. Accordingly, we will expand the corresponding description in the revised manuscript.
> RC3.3: "I anticipate that the packings were done with dry sands. How did you then add water afterwards to obtain uniform moisture contents in the packings?"
AR: The sand has not been purely dry, but exhibited a measurable soil water content as mentioned in the table. In order to minimize the change of water content while repacking the sand for the second experiment (due to evaporation, for instance), we removed and reinserted the sand within a few hours period. This way, we were able to keep the soil moisture rather constant. This will be clarified in the revised text.
> RC3.4: "Could there be environmental factors affecting the results?
AR: The tanks were located inside a large, air-conditioned hall, build and surrounded by concrete. This way we expect no influence external weather conditions. Between 5 and 50 meters, few scientists and cars may have been busy during the experiment period. This might be source of systematic uncertainty, but it is very hard to quantify. We will elaborate on the measurement setup in more detail.
> RC3.5: "l. 142: I suppose this should be for the loose experiment?
AR: Yes, this refers to the loose soil. We will clarify the text accordingly.
> RC3.6: "l. 155: In line 126 it is stated that the grain density in the simulations is 2.86 g/cm3 but now a value of 2.65 g/ cm3 is assumed? Any inconsistency here?"
AR: Line 126 refers to the density of the pure solid soil material (SiO2 + Al2O3), which is the default parameter in URANOS simulations. Line 155 refers to the particle density of quartz (SiO2), which is the accepted standard value for average soils. We agree that this is confusing, but this will only change the way how porosity and bulk density are converted by the two approaches, while principle results and conclusion of the study are not impacted. We will resolve this apparent inconsistency in the revised manuscript by adapting the porosity presentation and by pointing out that it is important to know about the two different parameters when comparing URANOS soil porosity with conventionally converted measurements.
> RC3.7: "Figure 1: In my printout, water and air have the same signature. Please make the two phases mutually discernable."
AR: We have optimized this figure to show maximum contrast and readability even for handicapped readers or grey-scale printers. Air has a pure white face color, solid soil is pure black, while water has a black dotted texture. The different signatures should also be clearly visible in black-and-white printouts. Please let us know how the contrast can be even further improved.
> RC3.7: "l. 164: It is 4 units of soil and 1 unit of water giving a ratio of 4. Please be consistent how you refer to this ratio throughout the manuscript."
AR: This is a typo, thanks for pointing this out. It will be fixed in the revised manuscript.
> RC3.7: "l. 165: equal 20 vol% porosity?"
AR: Yes, we will add "porosity" here.
> RC3.8: "l. 177: This sentence is contradictory to Figure 2. Intuitively I would think the sentence is right."
AR: This is an unfortunate typo, which also led to large confusion by Reviewer 2. It should read "it increases with increasing soil bulk density" and will be fixed.
> RC3.9: "l. 183: I am not sure I see a flat slope in Figure 3."
AR: We tried to say that the slope is constant, not flat, compared to the much more curved slopes of dry soils. We will correct the word in the revision.
> RC3.10: "Figure 3b: In Figure 3a the counts become higher for lower soil moisture contents but the opposite is the case in Figure 3b, which does not make sense.
AR: The plot is correct, as the low number of hydrogen and the high number of oxygen almost equally influence the neutron response. It is a notable example of the complex and non-linear relationship between neutrons and water content in the extreme dry regime. Please see section 4.2 for an elaborate discussion on this effect. We observed this opposing behaviour for soil moisture below 5%. To make it more visible, we split the figure into two panels.
> RC3.11: "Section 3.3 Evidence in the sand box experiment: There are some conflicting statements in this section regarding counts. Do the counts increase significantly (l. 193) or are they the same (l. 205)? What is the learning from the experiments given that some the controlling parameters are not constant?"
AR: This sentence refers to the topsoil experiment, where the deviation between the two solid-water ratio is very small. We apologize for the confusing statement. Since the experiment involved filling, unfilling, and compacting of 2 m³ of soil, it is almost impossible to exactly keep the soil moisture values constant. Nevertheless, we tried our best to reduce the risk for water evaporation during this process be refilling the box as fast as possible (within a few hours time). The learning from the experiments is very clear, as the first experiment provided a larger ratio change together with a larger change of neutron counts (confirming the N-bulk_density relationship), while the second experiment provided almost no change of the ratio together with almost no change of neutron counts (confirming our constant-ratio hypothesis).
> RC3.12: "l. 284-300: Please revise these lines as well as the figures. There are conflicting figure numbers and porosity numbers and some of the referred numbers of counts and soil moisture I do not see."
AR: The figure references are correct, but the writing is indeed a bit confusing. We will rephrase the text to describe the figures in a clearer way. Indeed, there is a typo in the schematic shown in Fig. 6, it should say "45%" porosity at the top and "35%" at the bottom.
> RC3.13: "l. 315: Impressive perhaps a bit exaggerated."
AR: We agree that this is speculation and will rephrase it in the revision.
> RC3.14: "l. 324-325: This is an important statement which perhaps should be emphasized even more."
AR: We believe that this statement does not require further discussion, but we will consider to slightly elaborate on this matter in the revised text. It has been made clear by Köhli et al. (2021) that the Koe function has been developed using different model assumptions than the Desilets equation. However, the performance of the Koe function regarding soil moisture and air humidity has been shown to be better than for the Desilets equation, particularly under extreme conditions. Hence, we base our study on the most recent state-of-the-art simulations by Köhli et al. (2021). In the present study we look at an additional aspect -- the influence of bulk density -- which has not been accounted for in previous studies. We acknowlegde that the usage of URANOS to investigate this effect is not as fair for the Des function as for Koe function, but it would not make sense to use deprecated model assumptions just to investigate the influence of bulk density on the Des function, while it has already been shown that the Des function has substantial issues on its own. Nevertheless, we decided to keep the comparison to the Desilets function in this study, in order to assess the potential impact on previous studies and their results using the Des function.
Citation: https://doi.org/10.5194/hess-2022-123-AC3
-
AC3: 'Reply on RC3', Martin Schrön, 24 Jul 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on hess-2022-123', Anonymous Referee #1, 10 May 2022
GENERAL COMMENTS
This paper comprehensively addressed the effects of soil bulk density on soil moisture estimation using cosmic-ray neutron sensing technique. The authors did a series of neutron simulations on a wide spectrum of soil bulk density and soil moisture. The simulation results were verified by conducting lab experiments of 4 scenarios sequentially. The effects of the bulk density were manifested by applying two widely used conversion equations on the simulation results. The authors also provide a correction approach for the CRNS community.
Overall, I think the authors did an acceptable job on simulation and lab experiment. However, the current version of the manuscript is not coherent or cohesive. There is plenty room for improvement in results presentation, content organization, and writing. Detailed suggestions and corrections are listed below.
SPECIFIC COMMENTS
Relationship to lattice water and organic matter
I am curious about if any effects of lattice water that has already been considered in the lattice water correction are considered again in the bulk density correction. The case with organic matter is similar. As mentioned in Line 267-269, mineral soils without any organic parts were considered. However, the low bulk density of soil is usually the consequence of high content of organic matter. In practice, if I apply both the organic matter correction and bulk density correction to my neutron counts, do I over-correct it?
Objectives and conclusions
The last two paragraphs of the introduction section need to be largely modified or rewritten. The hypothesis is only mentioned in this section but was not explained in detail how it is rejected/accepted and according to what criteria. To make the conclusion address the introduction, the objectives should be somewhat like 1) investigate the influence of bulk density in CRNS soil moisture measurement, 2) quantify the influence by examining the sensitivity of neutron counts through a large spectrum of soil bulk density, 3) assess the impact of bulk density on the two major conversion functions, and 4) develop a novel correction method/parameter for practical use.
Presentation of the results
The study provides substantial amount of simulation results. Many figures are clearly plotted and very informative. However, I feel a little awkward reading some of the figures. Based on the simulation and experiment design, I expected more figures like Fig. 5 with soil moisture as x-axis, and porosity contour lines, instead of the opposite, like Fig. 3. The presentation of Fig. 5 is similar to Fig. 6 in Zreda (2012) which is probably more common and easier to read in published literature.
The lab experiment is of great value to this study, which could potentially provide a great validation dataset for the simulation. I wonder if there is any experiment on varying soil moisture conducted on this setup. It would be great if there were more points on Fig. 3 or a separate figure of Nexp vs. Nsimu.
For Fig. 3 specifically, only 2 of the 4 experiment results were plotted in the figure. They were not colored with the soil moisture color scheme, which can barely inform the readers with the performance of the simulation/experiment. I suggest plotting a zoom-in panel of the four points with four ideal contour lines of the soil moisture content equal to the four measurements. The four points should also be colored with same color scheme.
Notation
The notations in this paper are generally clear but some selections are not very common. For dry bulk density, I think ρb is more common and ρs is often used as the density of solids instead of bulk density, which may cause some confusion.
For Section 3.1, “soil-water ratio” may be better named as “solid-water ratio”. You may assign this ratio with a symbol since it is also used in Table 1 and Section 3.3.
The correction factors
Some of the correction factors depend on soil moisture, which doesn’t really “go in line with the traditional correction factors for atmospheric changes…” (as mentioned in Line 337), since I think the correction factors should be independent of soil moisture. Line 374-375 mentioned a TDR sensor. Does it mean that a point-scale measurement or some rough estimate can also work here?
TECHNICAL CORRECTIONS
(not a complete list)
Line 28: in the soil -> In the soil
Line 98: Add definition of N0
Line 150: what is “DIN 18 125 T1 and T2”?
Line 176: Describe the y-axis. Is the “simulated neutron counts” with a. u. (arbitrary unit?) the same with the “neutron counts (normalized)”?
Line 208: section 2.2 -> Section 2.2 (many similar ones throughout the manuscript)
Line 209: Add “ in Koe. Function” before “(Eq. 8)”
Line 214: Extra space before “Eqs. 5”
Line 227-228: delete “above” and “less then”
Line 249: Fig. 3) -> Fig. 3a)
Line 254: (Fig.3) -> Fig. 3b)
Line 260-261: Any reference for the classification of bulk density? If the medium dense soil is between 1.6 and 1.8, this classification is not complete. For example, 1.5 g/cm3 doesn’t belong to any category.
Line 296: (6) -> Fig. 6
Line 340: section 2.2 -> Section 2.1
Line 356: based on; is only depend on -> only depends on
Line 358: the logic chain is broken between the two sentences
Line 370: extra “an”
Line 400: delete “small”
Figure 1: Instead of randomly giving 2 examples of the 7 simulations, describe the 7 simulations with the aid of Figure 1.
Table 1: Description of this table is not complete in the text. The column "differential" has different meanings across rows since the last row is in percentage. It is also mentioned as “deviation” in the text.
For Section 3.3, does it aim to verify/validate some of the simulation results? If so, consider renaming this section. The “previous section” (Line 206) is Section 3.2 or 3.1? If 3.1, consider switching/merging Section 3.2 and 3.3.
Citation: https://doi.org/10.5194/hess-2022-123-RC1 -
AC1: 'Reply on RC1', Martin Schrön, 24 Jul 2022
Dear Reviewer 1,
thank you very much for your positive review. We will briefly respond to your concerns in the interactive discussion. (RC=reviewer comment, AR=author response)
# Main comments
> RC1.0: "Overall, I think the authors did an acceptable job on simulation and lab experiment. However, the current version of the manuscript is not coherent or cohesive. There is plenty room for improvement in results presentation, content organization, and writing. Detailed suggestions and corrections are listed below."
AR: Thank you for your critical review and the suggestions to improve the manuscript. We will consider to restructure the text and to improve the writing throughout the revised manuscript.
##Specific comments
> RC1.1: "Relationship to lattice water and organic matter: I am curious about if any effects of lattice water that has already been considered in the lattice water correction are considered again in the bulk density correction. The case with organic matter is similar. As mentioned in Line 267-269, mineral soils without any organic parts were considered. However, the low bulk density of soil is usually the consequence of high content of organic matter. In practice, if I apply both the organic matter correction and bulk density correction to my neutron counts, do I over-correct it?"
AR: There is no risk of over-correction when correction for water-equivalent and bulk density are performed separately. As we have described in L43 and L269, lattice water and organic matter translate to an equivalent of water density, which is taken into account by the variable θ.
In general, epithermal neutrons are mainly sensitive to the density ratio between water and solid soil (as shown in section 2.1). They are rather insensitive to other properties of soils and molecular structure (see, e.g., Zreda et al. 2008, 2012 or Köhli et al. 2015). Hence, only those two components (soil density and hydrogen content) are to be considered when processing neutron data. In literature, soil density has not been considered yet (other than converting from gravimetric to volumetric), which is the motivation for the present study. Soil's hydrogen content has to be corrected for the hydrogen in lattice water as well as in organic material, in order to end up with a quantity representing mobile water/soil moisture only (see line 43). In the state of the art, this is done by simply substracting lattice water and organic water equivalent from the total water content measured by CRNS. The reviewer is right that one should correct for these variables, but this will only correct for hydrogen-related effects and not for any bulk density effects. Hence, these corrections can be treated independently.
On the other hand, we fully agree that low bulk densities are often correlated with high organic matter content. The value pairs of high porosity and low (effective) water content are therefore not very common in typical landscapes. In L269 we have already indicated this concept (Soils with bd = 1 g/cm³ have low abundance and commonly include a high organic carbon content which introduces an additional effect to θ."). We will make more clear that the soil moisture values used in the equations and in the figures comprise the total hydrogen content, i.e., mobile soil water + lattice water + organic water equivalent.
> RC1.2: "Objectives and conclusions: The last two paragraphs of the introduction section need to be largely modified or rewritten. The hypothesis is only mentioned in this section but was not explained in detail how it is rejected/accepted and according to what criteria. To make the conclusion address the introduction, the objectives should be somewhat like 1) investigate the influence of bulk density in CRNS soil moisture measurement, 2) quantify the influence by examining the sensitivity of neutron counts through a large spectrum of soil bulk density, 3) assess the impact of bulk density on the two major conversion functions, and 4) develop a novel correction method/parameter for practical use."
AR: Thank you for suggesting a clearer structure for the outline of the manuscript, we will consider adapting the last paragraphs of the introduction section accordingly. The reason why we initially decided to start with the negative hypothesis -- that neutrons are invariant against change of soil bulk density -- was that this has been the hitherto common assumption in CRNS literature. However, we understand that a more straight-forward description of the hypotheses would be more clear.
> RC1.3: "Presentation of the results: The study provides substantial amount of simulation results. Many figures are clearly plotted and very informative. However, I feel a little awkward reading some of the figures. Based on the simulation and experiment design, I expected more figures like Fig. 5 with soil moisture as x-axis, and porosity contour lines, instead of the opposite, like Fig. 3. The presentation of Fig. 5 is similar to Fig. 6 in Zreda (2012) which is probably more common and easier to read in published literature."
AR: Thanks for sharing your impression regarding the presented figures. We agree that neutrons-over-moisture figures are more common in CRNS literature, which is why we decided to plot Fig. 5 this way to illustrate the effect for a typical application. But the main body of the study is the investigation of the measurement variable (neutrons) in responds to the changing variable (soil bulk density). As it is common sense to plot the independent variable on the x-axis, and the dependent variable on the y-axis, we believe that neutron-over-density plots most clearly show the dependency of neutrons on bulk density, while the additional dependency on water content is of secondary importance.
> RC1.4: "The lab experiment is of great value to this study, which could potentially provide a great validation dataset for the simulation. I wonder if there is any experiment on varying soil moisture conducted on this setup. It would be great if there were more points on Fig. 3 or a separate figure of Nexp vs. Nsimu. For Fig. 3 specifically, only 2 of the 4 experiment results were plotted in the figure. They were not colored with the soil moisture color scheme, which can barely inform the readers with the performance of the simulation/experiment. I suggest plotting a zoom-in panel of the four points with four ideal contour lines of the soil moisture content equal to the four measurements. The four points should also be colored with same color scheme.
AR: Thank you for the suggestions. We will improve this figure visually and we will consider adding more data points if applicable.
> RC1.5: "Notation: The notations in this paper are generally clear but some selections are not very common. For dry bulk density, I think ρb is more common and ρs is often used as the density of solids instead of bulk density, which may cause some confusion. For Section 3.1, “soil-water ratio” may be better named as “solid-water ratio”. You may assign this ratio with a symbol since it is also used in Table 1 and Section 3.3."
AR: Thanks for the suggestions. From the physics point of view (section 2.1), dry soil bulk density is equivalent to the solid part of the soil, so we used abbreviations s (solid), w (water), and a (air) for the three main components, and we tried to be consistent with this notation in the remaining manuscript. Using ρb for soil bulk density would require a redefinition of ρs = ρb, which would be unnecessary and redundant. But we understand that it would be helpful to use the notations more common in soil science. Hence, we will reconsider the notation of dry soil bulk density in the revised manuscript.
> RC1.6: "The correction factors: Some of the correction factors depend on soil moisture, which doesn’t really “go in line with the traditional correction factors for atmospheric changes…” (as mentioned in Line 337), since I think the correction factors should be independent of soil moisture. Line 374-375 mentioned a TDR sensor. Does it mean that a point-scale measurement or some rough estimate can also work here?"
AR: We fully agree that the moisture-dependent correction is unconventional and may lead to a circular problem. The approach for correcting the road effect showed similar characteristics (Schrön et al. 2018). We tried very hard to find a solution that is independent of soil moisture, but unfortunately we cannot "work against the rules of nature". While moisture-dependent correction leads to the best performances, the good news is that moisture-free approaches also showed acceptable performance -- if the rough moisture regime is known (see the first two rows in Tables 2 and A1). So, yes, a very rough estimation of soil moisture is sufficient to achieve a decent correction performance. We will stress out this aspect more clearly in the revision.
# Technical corrections
AR: Thank you very much for your time to hint at technical issues and for making adequate suggestions. We will consider them thoroughly in the revised manuscript.
Citation: https://doi.org/10.5194/hess-2022-123-AC1
-
AC1: 'Reply on RC1', Martin Schrön, 24 Jul 2022
-
RC2: 'Comment on hess-2022-123', Anonymous Referee #2, 13 Jun 2022
Overall Comments
This paper presents an important topic for the calibration of CRNS sensors – whether or not there is a dependence on the soil bulk density. The issue has been somewhat neglected in the application of CRNS for soil moisture sensing, therefore this is an important study, especially for monitoring sites with tillage, or mobile CRNS applications, surveying across different soil types and bulk densities. The study is in the form of neutron transport simulations (using the URANOS model) and a dedicated lab experiment.
The presentation is reasonably good, with some additional details required. The English language is generally quite good, but should be improved. However, as it appears to me, there seems to be some major inconsistency between the theory and the presentation of results, and at times the text contradicts the plotted results. Attempting to untangle this is a major distraction from the message of the paper. Moreover, as presented currently I believe that the results are most likely incorrect, but will be corrected easily.
Detailed Points:
- The Discussion and Results present the increase of neutron counts with increasing bulk density (decreasing porosity) – whereas the theory (Equation 3) suggests that solid matter will scatter more efficiently than the much lower density air in the pore spaces, which implies to me that the neutron count should decrease with increasing bulk density (BD). Perhaps I am wrong, and misunderstand something here? But the text is also contradictory in places e.g. P.7, L.176-77:
“The highest neutron intensity can be achieved for very dry and dense soils, while it decreases with increasing soil bulk density (or decreasing porosity).”
If Fig.2 is plotted correctly, then this should at least be corrected to “….very dry and low density [loose] soils…”
However, I have doubts whether the plots are correct? For Fig.2 & Fig.3 (and subsequent plots with BD/porosity x-axis) – I believe that all the x-axes labels are reversed? Making this assumption, the results would make physical sense, otherwise, as above the results do not match (or are opposite to) the theory presented, as I understand it.
The later text does match the plots, so I am really left wondering if all the discussion and presentation of results has to be revised, as well as the plots… or else explain what I am missing and why we should expect higher neutron counts with higher BD?
- The paper is written in the future tense (e.g. P.2, L.54-55), this not the conventional way to report scientific experiments.
- Take more care with the use of tenses, and prepositions e.g. Title change “influence to soil moisture” to “influence on soil moisture”. English should be generally improved and errors removed.
- Section 2.4 Experimental Concept: Where are the tanks located? - how did you control the surrounding environment? Could there have been other non-constant hydrogen pools in the surrounding 100 to 200m?
- 5, L.139-L.140 specify the CRNS detector type (make & model). Why only a 60 minute counting period? Why not longer to improve the count statistics?
- Some figure captions (Fig.2 etc.) have the wrong units for BD – should be g/cm^3 not g/m^3.
- 9, L.191 were the neutron counts first corrected in any way? For example, for changes in atmospheric pressure, air humidity, incoming neutron intensity? Otherwise, neutron counts made at different times cannot be directly compared!
- Table 1. Are the neutron counts assigned to the correct porosities? Whilst data here matches Fig.3, it does not agree with the theory (Eq. 3)
- 11, L.240-242 Is this the right way round? Here both BD and SM vary - so which has the more dominant effect on penetration depth?
- 13, L.293 understimate (?) --> increased porosity should increase counts! Leading to underestimate of SM??.... and Fig.5 Why don't counts increase with porosity? I don't think this is correct?
- Summary & Conclusions: (2.) “On average, neutron count rates decrease by −1% for every +10% increase in porosity”. This appears contrary to Eq.3 ? I would expect an increase in count rate. As per Point 1 above, much of this text needs revising if indeed the plot x-axes are reversed.
Citation: https://doi.org/10.5194/hess-2022-123-RC2 -
AC2: 'Reply on RC2', Martin Schrön, 24 Jul 2022
Dear Reviewer 2,
thank you very much for your critical and positive review. We will briefly respond to your concerns in the interactive discussion. (RC=reviewer comment, AR=author response)
# Overall comments
> RC2.1: "The presentation is reasonably good, with some additional details required. The English language is generally quite good, but should be improved. However, as it appears to me, there seems to be some major inconsistency between the theory and the presentation of results, and at times the text contradicts the plotted results. Attempting to untangle this is a major distraction from the message of the paper. Moreover, as presented currently I believe that the results are most likely incorrect, but will be corrected easily."
AR: Thank you for sharing your impression. We believe that the reviewer misinterpreted the slope of the curve of equation 3, which led to an apparent inconsistency with the rest of the text and results. In the detailed response below we will try to solve this misconception, proving that the theory is indeed consistent with the simulations.
Large parts of this review refer to apparent inconsistencies of the figures, texts, and conclusions with equation 3. These comments are seemingly all based on an initial misconception, which can be quickly solved (see next section). However, we admit that a typo in line 177 ("decreasing" instead of "increasing") may have led to this substantial confusion. We apologise for this mistake will be fix the sentence in the revision.
> RC2.1: "The Discussion and Results present the increase of neutron counts with increasing bulk density (decreasing porosity) – whereas the theory (Equation 3) suggests that solid matter will scatter more efficiently than the much lower density air in the pore spaces, which implies to me that the neutron count should decrease with increasing bulk density (BD). Perhaps I am wrong, and misunderstand something here?"
AR: Your are referring to equation 3 (or the simplified equation 4), where N is proportional to a combination of cross sections, densities, and energy loss quantifiers. Here, the slope of the curve (N over s) highly depends on the weighting factors ξs and ξw in the denominator. It is 1 for ξs = ξw, <1 for ξs > ξw, and >1 for ξs < ξw. According to Köhli et al. 2021, the variable xi represents the energy loss per collision, and can be estimated with ξ=2/(A+1), where A is the atomic mass number. For hydrogen, ξ(A=1)=1, and for Silicon, ξ(A=28)=0.07. Hence, ξs << ξw.
For a quick illustration of how this influences N, let a=s*Σs and b=w*Σw, then N ~ (a+b)/(a*ξs + b*ξw). With ξs << ξw this reduces roughly to N~(a+b)/b = 1+a/b, i.e., N ~ s. Using rough values for all the variables, the following link shows this relationship plotted: https://tinyurl.com/neutronsoliddensityplot
Since the theory indeed confirms our simulation results, we can be assured that the plots and conclusions are correct. However, we realize that better understanding of the theoretical part is critical for the readers, so we will consider improving the explanations and guidance on interpreting the theoretical concepts.
> RC2.2: "But the text is also contradictory in places e.g. P.7, L.176-77:
“The highest neutron intensity can be achieved for very dry and dense soils, while it decreases with increasing soil bulk density (or decreasing porosity)."AR: You are right that this sentence contradicts itself. We meant, that neutron intensity *increases* with increasing soil bulk density. Sorry for the confusion, it will be changed in the revision.
> RC2.3: "However, I have doubts whether the plots are correct? For Fig.2 & Fig.3 (and subsequent plots with BD/porosity x-axis) – I believe that all the x-axes labels are reversed? Making this assumption, the results would make physical sense, otherwise, as above the results do not match (or are opposite to) the theory presented, as I understand it.
The later text does match the plots, so I am really left wondering if all the discussion and presentation of results has to be revised, as well as the plots… or else explain what I am missing and why we should expect higher neutron counts with higher BD?AR: The plots are correct and do confirm the theory (see above). The text will be double-checked for potentially confusing formulations.
> RC2.4: "The paper is written in the future tense (e.g. P.2, L.54-55), this not the conventional way to report scientific experiments.
Take more care with the use of tenses, and prepositions e.g. Title change “influence to soil moisture” to “influence on soil moisture”. English should be generally improved and errors removed."AR: Thank you, we will improve the writing upon revision of the manuscript.
> RC2.5: "Section 2.4 Experimental Concept: Where are the tanks located? - how did you control the surrounding environment? Could there have been other non-constant hydrogen pools in the surrounding 100 to 200m?
AR: The tanks were located inside a large, air-conditioned hall, build and surrounded by concrete. This way we expect no influence external weather conditions. Between 5 and 50 meters, scientists and cars have been present but with only minor rearrangement during the experiment period. This might be a valid source of systematic uncertainty, but it is constant and very hard to quantify. We will elaborate on the measurement setup in more detail.
> RC2.6: "5, L.139-L.140 specify the CRNS detector type (make & model). Why only a 60 minute counting period? Why not longer to improve the count statistics?
AR: The detector has been specified as the one used for roving in Schrön et al. 2018, with roughly 10x more count rate than CRS1000 detectors. The measurement period is short because we tryed to minimize the risk of water loss due to evaporation during the day. Our main goal was to keep the soil water content constant between the two experiments (dense and loose), so we removed the soil and repacked it within one day, leaving only about 60 minutes time for each of the two measurements. We will better describe this approach in the text.
> RC2.7: "L.191 were the neutron counts first corrected in any way? For example, for changes in atmospheric pressure, air humidity, incoming neutron intensity? Otherwise, neutron counts made at different times cannot be directly compared!"
AR: Yes, we corrected the neutron counts according to standard procedures. We will add this information to the revised manuscript.
> RC2.8: "Table 1. Are the neutron counts assigned to the correct porosities? Whilst data here matches Fig.3, it does not agree with the theory (Eq. 3)"
AR: We are glad to report that the data is consistent with equation 3, see the explanation above.
> RC2.9: "L.240-242 Is this the right way round? Here both BD and SM vary - so which has the more dominant effect on penetration depth?"
AR: We agree that this example is challenging to interpret. We will provide a more clear example that shows the different contributions of soil moisture and bulk density on the penetration depth. However, this is just a side note and based on already published literature (see D86 in Schrön et al. 2017).
> RC2.10: "13, L.293 understimate (?) --> increased porosity should increase counts! Leading to underestimate of SM??.... and Fig.5 Why don't counts increase with porosity? I don't think this is correct?
Summary & Conclusions: (2.) “On average, neutron count rates decrease by −1% for every +10% increase in porosity”. This appears contrary to Eq.3 ? I would expect an increase in count rate. As per Point 1 above, much of this text needs revising if indeed the plot x-axes are reversed."AR: Also here the text is consistent with the results and theory (see above).
Citation: https://doi.org/10.5194/hess-2022-123-AC2
-
EC1: 'Comment on hess-2022-123 - Start interacting', Nunzio Romano, 21 Jun 2022
While waiting for another reviewer's evaluation, I suggest you should start providing some preliminary responses to the comments received so far. This helps feed the discussion step of the journal.
Citation: https://doi.org/10.5194/hess-2022-123-EC1 -
RC3: 'Comment on hess-2022-123', Anonymous Referee #3, 26 Jun 2022
The manuscript reports a numerical and experimental study of the effect of soil bulk density on cosmic ray sensing (CRNS). Such an impact is important if the land surface is subject to farming practices and for applications of the CRNS method for mapping soil moisture across heterogeneous landscapes.
I believe that the impact of soil bulk density is an impact factor that needs to be considered at least for some applications when converting neutron counts into soil moisture and as such, the study is highly relevant and important and ultimately warrants publication. However, the manuscript is not ready for publication in its present form as far as the analysis, presentation and language are concerned. The authors need to make a thorough effort in improving the text and make it consistent both internally and in relation to the presented figures. During my reading of the manuscript, I noted language issues at several places but I will only mention some of them below.
Comments:
l. 50: Perhaps a bit unusual to start out with the hypothesis that that neutron intensity is unaffected by soil bulk density. The whole idea of the manuscript is to document that the opposite is the case, so it would be more intuitively to state that the objective of the work is to analyze and quantify the impact of soil bulk density on neutron counts and how to take this impact into consideration to obtain reliable estimates of soil moisture.
Section 2.4 Experimental concept: I believe that the reported experiments are very interesting in order to gain experimental evidence of the impact of soil bulk density. The experiments are indeed a comprehensive endeavor and I would like to know a bit more details on the setup perhaps by an accompanying figure, which also shows the position of the CRNS detector.
I anticipate that the packings were done with dry sands. How did you then add water afterwards to obtain uniform moisture contents in the packings?
Could there be environmental factors affecting the results?
l. 142: I suppose this should be for the loose experiment?
l. 155: In line 126 it is stated that the grain density in the simulations is 2.86 g/cm3 but now a value of 2.65 g/ cm3 is assumed? Any inconsistency here?
Figure 1: In my printout, water and air have the same signature. Please make the two phases mutually discernable.
l. 164: It is 4 units of soil and 1 unit of water giving a ratio of 4. Please be consistent how you refer to this ratio throughout the manuscript.
l. 165: equal 20 vol% porosity?
l. 177: This sentence is contradictory to Figure 2. Intuitively I would think the sentence is right.
l. 183: I am not sure I see a flat slope in Figure 3.
l. 186: Again, I would not call it a flat curve.
Figure 3b: In Figure 3a the counts become higher for lower soil moisture contents but the opposite is the case in Figure 3b, which does not make sense.
Section 3.3 Evidence in the sand box experiment: There are some conflicting statements in this section regarding counts. Do the counts increase significantly (l. 193) or are they the same (l. 205)? What is the learning from the experiments given that some the controlling parameters are not constant?
l. 284-300: Please revise these lines as well as the figures. There are conflicting figure numbers and porosity numbers and some of the referred numbers of counts and soil moisture I do not see.
l. 315: Impressive perhaps a bit exaggerated.
l. 324-325: This is an important statement which perhaps should be emphasized even more.
Citation: https://doi.org/10.5194/hess-2022-123-RC3 -
AC3: 'Reply on RC3', Martin Schrön, 24 Jul 2022
Dear Reviewer 3,
thank you very much for your critical and detailed review. We will briefly respond to your concerns in the interactive discussion. (RC=reviewer comment, AR=author response)
# Comments
> RC3.0: "I believe that the impact of soil bulk density is an impact factor that needs to be considered at least for some applications when converting neutron counts into soil moisture and as such, the study is highly relevant and important and ultimately warrants publication. However, the manuscript is not ready for publication in its present form as far as the analysis, presentation and language are concerned. The authors need to make a thorough effort in improving the text and make it consistent both internally and in relation to the presented figures. During my reading of the manuscript, I noted language issues at several places but I will only mention some of them below."
AR: Thanks for sharing your impression on the manuscript. We thank the reviewer for pointing out some language and writing issues, which will be solved during the revision. However, we have not been able to identify issues with the analysis based on this review.
> RC3.1: "l. 50: Perhaps a bit unusual to start out with the hypothesis that that neutron intensity is unaffected by soil bulk density. The whole idea of the manuscript is to document that the opposite is the case, so it would be more intuitively to state that the objective of the work is to analyze and quantify the impact of soil bulk density on neutron counts and how to take this impact into consideration to obtain reliable estimates of soil moisture.
AR: We agree and will reformulate the last paragraphs of the introduction according to the suggestions from Reviewer 1. The reason why we initially decided to start with the negative hypothesis (that neutrons are invariant against change of soil bulk density) was that this has been the hitherto common assumption in CRNS literature. However, we understand that a more straight-forward description of the hypotheses would be more comprehensive.
> RC3.2: "Section 2.4 Experimental concept: I believe that the reported experiments are very interesting in order to gain experimental evidence of the impact of soil bulk density. The experiments are indeed a comprehensive endeavor and I would like to know a bit more details on the setup perhaps by an accompanying figure, which also shows the position of the CRNS detector."
AR: Thanks for the suggestion, also the other reviewers asked for a more elaborate description of the experiment. Initially, we did not plan to elaborate more on the experiment, since the main focus of this manuscript is the theory (and simulations). But we agree that the lab experiment is a unique setup which deserves more detailed explanations. Accordingly, we will expand the corresponding description in the revised manuscript.
> RC3.3: "I anticipate that the packings were done with dry sands. How did you then add water afterwards to obtain uniform moisture contents in the packings?"
AR: The sand has not been purely dry, but exhibited a measurable soil water content as mentioned in the table. In order to minimize the change of water content while repacking the sand for the second experiment (due to evaporation, for instance), we removed and reinserted the sand within a few hours period. This way, we were able to keep the soil moisture rather constant. This will be clarified in the revised text.
> RC3.4: "Could there be environmental factors affecting the results?
AR: The tanks were located inside a large, air-conditioned hall, build and surrounded by concrete. This way we expect no influence external weather conditions. Between 5 and 50 meters, few scientists and cars may have been busy during the experiment period. This might be source of systematic uncertainty, but it is very hard to quantify. We will elaborate on the measurement setup in more detail.
> RC3.5: "l. 142: I suppose this should be for the loose experiment?
AR: Yes, this refers to the loose soil. We will clarify the text accordingly.
> RC3.6: "l. 155: In line 126 it is stated that the grain density in the simulations is 2.86 g/cm3 but now a value of 2.65 g/ cm3 is assumed? Any inconsistency here?"
AR: Line 126 refers to the density of the pure solid soil material (SiO2 + Al2O3), which is the default parameter in URANOS simulations. Line 155 refers to the particle density of quartz (SiO2), which is the accepted standard value for average soils. We agree that this is confusing, but this will only change the way how porosity and bulk density are converted by the two approaches, while principle results and conclusion of the study are not impacted. We will resolve this apparent inconsistency in the revised manuscript by adapting the porosity presentation and by pointing out that it is important to know about the two different parameters when comparing URANOS soil porosity with conventionally converted measurements.
> RC3.7: "Figure 1: In my printout, water and air have the same signature. Please make the two phases mutually discernable."
AR: We have optimized this figure to show maximum contrast and readability even for handicapped readers or grey-scale printers. Air has a pure white face color, solid soil is pure black, while water has a black dotted texture. The different signatures should also be clearly visible in black-and-white printouts. Please let us know how the contrast can be even further improved.
> RC3.7: "l. 164: It is 4 units of soil and 1 unit of water giving a ratio of 4. Please be consistent how you refer to this ratio throughout the manuscript."
AR: This is a typo, thanks for pointing this out. It will be fixed in the revised manuscript.
> RC3.7: "l. 165: equal 20 vol% porosity?"
AR: Yes, we will add "porosity" here.
> RC3.8: "l. 177: This sentence is contradictory to Figure 2. Intuitively I would think the sentence is right."
AR: This is an unfortunate typo, which also led to large confusion by Reviewer 2. It should read "it increases with increasing soil bulk density" and will be fixed.
> RC3.9: "l. 183: I am not sure I see a flat slope in Figure 3."
AR: We tried to say that the slope is constant, not flat, compared to the much more curved slopes of dry soils. We will correct the word in the revision.
> RC3.10: "Figure 3b: In Figure 3a the counts become higher for lower soil moisture contents but the opposite is the case in Figure 3b, which does not make sense.
AR: The plot is correct, as the low number of hydrogen and the high number of oxygen almost equally influence the neutron response. It is a notable example of the complex and non-linear relationship between neutrons and water content in the extreme dry regime. Please see section 4.2 for an elaborate discussion on this effect. We observed this opposing behaviour for soil moisture below 5%. To make it more visible, we split the figure into two panels.
> RC3.11: "Section 3.3 Evidence in the sand box experiment: There are some conflicting statements in this section regarding counts. Do the counts increase significantly (l. 193) or are they the same (l. 205)? What is the learning from the experiments given that some the controlling parameters are not constant?"
AR: This sentence refers to the topsoil experiment, where the deviation between the two solid-water ratio is very small. We apologize for the confusing statement. Since the experiment involved filling, unfilling, and compacting of 2 m³ of soil, it is almost impossible to exactly keep the soil moisture values constant. Nevertheless, we tried our best to reduce the risk for water evaporation during this process be refilling the box as fast as possible (within a few hours time). The learning from the experiments is very clear, as the first experiment provided a larger ratio change together with a larger change of neutron counts (confirming the N-bulk_density relationship), while the second experiment provided almost no change of the ratio together with almost no change of neutron counts (confirming our constant-ratio hypothesis).
> RC3.12: "l. 284-300: Please revise these lines as well as the figures. There are conflicting figure numbers and porosity numbers and some of the referred numbers of counts and soil moisture I do not see."
AR: The figure references are correct, but the writing is indeed a bit confusing. We will rephrase the text to describe the figures in a clearer way. Indeed, there is a typo in the schematic shown in Fig. 6, it should say "45%" porosity at the top and "35%" at the bottom.
> RC3.13: "l. 315: Impressive perhaps a bit exaggerated."
AR: We agree that this is speculation and will rephrase it in the revision.
> RC3.14: "l. 324-325: This is an important statement which perhaps should be emphasized even more."
AR: We believe that this statement does not require further discussion, but we will consider to slightly elaborate on this matter in the revised text. It has been made clear by Köhli et al. (2021) that the Koe function has been developed using different model assumptions than the Desilets equation. However, the performance of the Koe function regarding soil moisture and air humidity has been shown to be better than for the Desilets equation, particularly under extreme conditions. Hence, we base our study on the most recent state-of-the-art simulations by Köhli et al. (2021). In the present study we look at an additional aspect -- the influence of bulk density -- which has not been accounted for in previous studies. We acknowlegde that the usage of URANOS to investigate this effect is not as fair for the Des function as for Koe function, but it would not make sense to use deprecated model assumptions just to investigate the influence of bulk density on the Des function, while it has already been shown that the Des function has substantial issues on its own. Nevertheless, we decided to keep the comparison to the Desilets function in this study, in order to assess the potential impact on previous studies and their results using the Des function.
Citation: https://doi.org/10.5194/hess-2022-123-AC3
-
AC3: 'Reply on RC3', Martin Schrön, 24 Jul 2022
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,171 | 382 | 46 | 1,599 | 71 | 30 | 34 |
- HTML: 1,171
- PDF: 382
- XML: 46
- Total: 1,599
- Supplement: 71
- BibTeX: 30
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Mandy Kasner
Steffen Zacharias
This preprint has been withdrawn.
- Preprint
(474 KB) - Metadata XML
-
Supplement
(9 KB) - BibTeX
- EndNote