Alum and Gypsum Amendments Decrease Phosphorus Losses from Soil Monoliths to Overlying Floodwater under Simulated Snowmelt Flooding

Round 1

Reviewer 1 Report

There needs to be a lot more coverage and citations in the intro with regard to many studies that have been conducted on the use of these amendments on soils for P.  I realize that this current study is on flooded soils, but the non-flooded soils need to be covered.

The results section is much too long.  Each section within the results can be reduced to a paragraph.  You write extensively about various interactions between variables, but then never discuss why that is, or use that information in the context of your objectives.  If it is not being discussed, delete it. 

While the conclusions stated in the conclusions section are fine and supported by the results, there are serious flaws in the discussion and faulty assertions.  See details on that, below.  But in general, correlation does not equal causation.  The authors make this mistake several times.  You must have a real “physical reality” in explaining a correlation.  This makes the manuscript very weak.  In addition, the use of the thermodynamic model is overly-relied on, and not completely used correctly.

From lines 169-171: Why not permanently install the reference electrode as well?  why leave the redox probe installed but not the reference since both are needed for the measurement?

The assumptions about gypsum being present at a certain concentration and about the DOM and DOC need to be justified.  Lines 206-208.

The discussion on the interaction between P and organic matter is rather flawed:  Lines 484-488:

“Previous studies have also shown that gypsum amendment decreased the solubility of organic P in soils, which was attributed to the greater stability of organic matter complexes due to increased ionic strength and Ca concentrations [49; 57]. Our results, showing lower DOC concentrations in gypsum and alum amended treatments of both Dencross 2 and Lakeland soils, corroborate this assertion.”

You need more evidence and clarity to make such an assertion.  First, what forms of organic P are you referring to?  biomolecules that contain P, or P sorbed onto organic compounds?  The former requires biological interaction to mineralize the P.  The latter (DOC) does not effectively adsorb P.  humic acid and fulvic acid do not sorb P, which is well known and observed in that common greenhouse media does not sorb P and therefore require high P concentrations in fertigation.  Those organic compounds do not "complex" P as they do not have the proper functional groups.  Recall that DOC is able to complex trace metals because of many non-polar functional groups, but that is not effective for phosphate.  In fact, the literature is full of studies that illustrate how DOC actually competes with P, often reducing P sorption because both are complexing agents.  You have no direct evidence of this, and correlation does not equal causation.  Please delete as it is very misleading.

 From line 524: “With decreasing pe + pH with flooding time, porewater concentrations of Ca and Mg generally increased suggesting dissolution of Ca and Mg phosphates”. 

In soils, Ca and Mg concentrations typically increase with decreasing pH.  So your statement here contradicts what you previously stated about how pH was steady or slightly decreased with flooding (336-339).  Therefore, the increasing Ca and Mg concentrations with pe + pH must be due to the changes in pe.  Yet, pe has little to no effect on Ca and Mg minerals, so there must be another explanation.   This is another example of assuming that correlation equals causation.  Just because decreasing pe is correlated with increasing Ca and Mg concentrations, does not mean that it is a causal relationship.  You need to establish some sort of a real physio-chemical process to that, otherwise it is meaningless and rather misleading.  For example, with regard to Mn and Fe, you can easily attribute such a relationship to a real process that is occurring.  You need to do the same for Ca and Mg and not simply assume that a correlation is equal to causation. 

Instead, what is more likely the reason for increased Ca and Mg concentrations, since pH was not decreasing, is that the added TIME allowed for dissolution of Ca and Mg minerals.  Remember that dissolution takes time, even if a thermodynamic calculation predicts dissolution, it may require a long amount of time for it to occur.  Indeed, soil generally release more cations with increased time of contact with a water solution; this has been well established.  It takes time for the soil to equilibrate to the surrounding solution.  Another strong and logical possibility is that Mg and Ca mineral solubility increased with decreasing concentration of CO₂, since respiration is decreasing under flooded conditions.  Keep in mind that a decrease in the concentration of CO₂ will increase the solubility of Ca and Mg minerals, and that does take time to develop.  Again, time is a critical factor. 

Second, it is unlikely that the increased Ca and Mg in solution is from dissolving Ca and Mg phosphate minerals.  First, you have no evidence for that.  The use of the SI do not support this.  SI simply tells you if there is POTENTIAL for a mineral to be precipitated or not.  It does not take into account time.  If it is precipitated, it is precipitated.  It can't be more or less precipitated.  SI is either positive, zero, or negative.  Second, the amount of Ca and Mg released from a dissolving Ca and Mg phosphates will be tiny compared to the giant increase in Ca and Mg that you are seeing.  Such high concentrations of Ca and Mg in solution are precisely what causes Ca and Mg phosphates to precipitate with P.  In other words, there is another source of Ca and Mg that is supplying the solution, and this in turn is what controls the solubility of Ca and Mg phosphates.  Consider the following:

5Ca²⁺_(aq) + 3H₂PO₄^-_(aq) + H₂O ßà Ca₅(PO₄)₃OH_(s) (hydroxyapatite)+ 7H⁺  

CaCO₃ ßà Ca²⁺ + CO₃^2-

The dissolution of the second reaction creates more Ca²⁺ in solution, and this feeds the first reaction to actually precipitate more Ca phosphate.  With carbonate in equilibrium with CO₂, note also that decreased CO₂ will cause the Ca (and Mg) minerals to dissolved more.

Regardless, if we accept your underlying assumptions and for the best-case scenario, assume that all of the increased P concentration in solution is coming from a Ca phosphate mineral, how much Ca would be released?  If you look at the unamended soil for example, you have about 20 mg/L Ca at pe+pH of 15, which occurs around day 7.  At pe+pH of 9, which occurs at around day 49-56, you get an increase to around 150 mg Ca/L.  Over that same time period, the max increase in P concentration is only 1 mg P/L.  Consider that hydroxyapatite gives you max Ca:P at 10 moles of Ca per 6 moles of P, which is 2.15 mg Ca per mg P.  Thus, an increase in 1 mg P/L coming from a dissolving hydroxyapatite mineral would only increase the dissolved Ca concentration by 2.15 mg/L, which is not nearly enough to account for the 130 mg/L increase that you observed for the unamended soil.  Thus, there must be a different and larger source of Ca and Mg other than Ca and Mg phosphates.  In addition, that 1 mg P/L increase is also coming from other minerals, such as the Fe minerals that are reduced and dissolved.  Bottom line is that this entire discussion on Ca and Mg phosphates dissolution, and dissolution of Ca and Mg due to pe + pH is baseless and illogical.  Please delete it because it is misleading.  In fact, please delete the figures that show changes in Ca and Mg concentrations with pe+pH.  You have no processes to base any of it on, only correlations, and those correlations are meaningless without being grounded in a physical reality.  Redox does not affect Ca, as the thermodynamic constants from such redox reactions are very unfavorable.  For example, From Lindsay, 1979

Ca²⁺ + 2e- ßà Ca⁰  log K = -97.16

pe = -48.58 + 0.5 log Ca²⁺

He states that “an extremely low pe is required for Ca0 to be stable”.  Notice that you do not see Ca or Mg listed in the hierarchy of chemicals that become reduced in soils (Nitrate, Fe, Mn, sulfate, etc..).  Compare the pe required for Fe and Mn.  Thus, you have no physio-chemical basis for drawing connections or conclusions with redox and Ca/Mg.  Be very careful with correlations as they can be very misleading.

Instead of showing Ca and Mg concentrations in relation to pe+pH, shown them in relation to time.  This would actually be useful to the reader.

Similarly, you conclude in line 553 that the dissolution of Ca-P increased P concentrations at the beginning of flooding.  But you have no evidence for this.  Again, the SI is just a guide, and it only gives you an idea about the potential for a mineral to be present or not.  You indicate in Table 1 that the system is saturated in Ca and P for the unamended and gypsum soils, meaning that there is potential for precipitation (i.e. a positive SI)  It is positive and it stays positive.  You cannot make conclusions from an SI value being positive and decreasing, yet still being positive.  It is either positive, zero, or negative.  Not only that, but even if that was valid, you have no ability to compare values in the table.   Consider that there are other sources of P in the soil that are not considered with the Minteq model you are using.  P bound to minerals by ligand exchange, and also bound to the outer sphere via AEC.  Ultimately, your conclusion is clearly unfounded.  This also contradicts what you stated earlier about how Ca and Mg phosphate dissolution is increasing with flood time (line 524).

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Title: Alum and gypsum amendments decrease phosphorus losses from soil monoliths to overlying floodwater under simulated snowmelt flooding

Abstract:

I find this section well written as it captures the entire research done.

Introduction:

I also find this section well written.

Materials and Methods:

This section is also well structured and the science is valid. I only suggest adding some diagrams, photos, maps

Results: I find the results well-presented and structured

Discussion: The study is fairly well discussed but requires some revision. The discussion was mainly focused on the current study with very little clear comparisons with other similar studies.

Conclusion: The manuscript is fairly well concluded.

Although, the study is interesting and will appeal to readers, I suggest the comments raised (and from all reviewers) are duly addressed to make it more comprehensible and concise to bring the quality to a publishable level.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Authors,

Congratulations for the huge work done and for the nice results.

My first feeling was that this article includes at least 3 articles.

And I found that you already had some related articles published.

Still, I think this amount of information is enough for at least 2 articles. The text is 23 pages long and there are 9 extra pages in the supplementary part.

Please see some comments below.

 

Introduction

I am not from a cold climatic region, so maybe this is why I need some clarification. It is a minor issue, one might consider that we are just going into details that might not have even minor concern but still, there are 2, (I think) distinct phenomenon that the readers can read here:

1. Lines 38–39. „In cold climatic regions, spring snowmelt runoff is a dominant pathway of P losses from soils . . . ” and
2. Line 39–40. „During the snowmelt period in the Canadian prairies, agricultural fields remain flooded . . .”

So, my clarifying question is if the runoff from the field was creating the flooding at the bottom of the slope or the snowmelt runoff is “only” the source of the excess water in rivers and the rivers are flooding and causing the floods over the agricultural fields?

I miss the “Visual MINTEQ modeling” part. Who was using the model, where and for what purposes?

Results

Soil characterization

Some authors’ former article includes the same soils and their evaluation. Would not it be enough to refer to this article and only add the new data that was not published (e.g. Olsen P) yet? Furthermore, it would be nice to clarify some details, e.g. are the soils “strongly” or “weakly to moderately” calcareous?

Environmental Pollution Volume 287, 15 October 2021, 117619   Phosphorus mobilization in unamended and magnesium sulfate-amended soil monoliths under simulated snowmelt flooding   Udaya W.A. Vitharana, Darshani Kumaragamage, B.L.W.K. Balasooriya, Doug Goltz	Water – recent paper under review   Alum and gypsum amendments decrease phosphorus losses from soil monoliths to overlying floodwater under simulated snowmelt flooding   Darshani Kumaragamage, Chamara S. Weerasekara, Madelynn Perry, Wole O. Akinremi and Doug Goltz
Characterization of soils/Soil characterization
2.2.	3.1.
The CCE values ranged from 89 to 167 g kg−1 indicating that all soils are strongly calcareous.	These soils were weakly to moderately calcareous with CCE values between 90 and 146 g kg-1.
Can you explain the slight change in the CCE values range?
Calcareous nature of soils was further reflected by high extractable Ca (>137.3 mg kg−1) and Mg (>95.1 mg kg−1) contents.	Mehlich extractable cations were in the range of 1378 to 3872 mg kg-1 for Ca, 956 -2008 mg kg-1 for Mg
Can you explain the tremendous (tenfold) difference in magnitude between the two evaluations?

Furthermore, I wonder if the CEC values measured by the introduced methodology were correct? How can you evaluate CEC values if you calculate it by the mentioned extractable cations (M3 method)? Do you get the same CEC values?

I mean, some criticism should be applied when you have a total amount of 5842 mg/kg of Ca+Mg+Fe+Mn+Al in case of Dencross 1 soil and 5197 mg/kg of Ca+Mg+Fe+Mn+Al in case of Dencross 2 soil, so Dencross 2 soil had appr. 11% less than Dencross 1, while the CEC value measured by the Na+ method resulted 45 for Dencross 1 and almost twice as much, 86 for Dencross 2, so the less amount of Ca+Mg+Fe+Mn+Al had the higher CEC value. I guess there were not huge differences in the amounts of other cations, or were there? What they might be and how do they affect the results? And if the CEC values measured by the Na+ method are not correct, how is this affect the results? Or I am off the track and besides 3700+ and 3800+ mg/kg Ca and 1500+ and 1100+ mg/kg Mg there were other influencing cations in these soils?

I wonder what the parent rocks were?

Another issue is the amount of P measured by the Olsen and the Mehlich 3 methods. The difference is from 117 to 350 % favoring M3 method. How did you take these differences into account? Did you use these data at all? What might be the result of these differences on your evaluation of P losses?

 

3.2. subchapter – Figure 1.

I understand the aim of maximize the size of the figure, so it can be easily read but it is reducing the chances of the visibility of the comparison of the data, there are maximum values of 1, 3, 4 and 6, sometimes it is not even reasonable. I suggest using 4 as maximum on the Y-axis, it would help presenting the differences between the soils.

3.2. subchapter – Figure 2.

I also suggest using the same maximum (maybe 1.5 would be nice) for the Y-axis, since, I guess, it is important to compare the behavior of the different soils!

 

Overall “evaluation” of the results

The description of the results are very spacious and often very difficult to follow, especially with the data published in the supplementary part and with the data not published.

I miss some evaluation of the data, e.g. Line 284 and 287 (see my comment). I think these are very important details and the chances of their evaluation were left behind (neglected).

These two lines are only examples (we could go on for Line 293, 297, 413, etc.). I have a feeling that a lot of results were “just” written and the evaluation is missing. And since this is the results section, followed by the discussion, there was no space left for the evaluation. I think. And I miss it a lot! Especially because there are results that could be expected, so we can say that yes, of course pH decreased because . . . and DRP decreased or increased her and there but what could be the reason?

I think that these small details, giving at least “some” hints for the readers, for the farmers or for decision makers, what to expect with the different soil series or soil properties might increase the value of the manuscript.

 

Discussion

The discussion is very detailed!

Lines 559–564. You wrote: “Gypsum-amended treatments which maintained a higher Eh than other treatments, favored the formation of Fe(III) hydr(oxides) such as hematite (α-Fe2O3), lepidocrocite (γ-FeOOH), goethite (α-FeOOH) at 56 DAF. Formation of these P-sorbing mineral species may facilitate re-adsorption of released P, as evidenced by the slight decrease, or relatively stable DRP concentrations towards the latter stages of flooding.”

I do not see any introduction of this section with reference to the minerals (hematite, lepidocrocite, goethite, etc.). I see the modelling part in materials and methods and also Suppl. Table 8. explains that Visual MINTEQ modeling was used to predict these minerals but I do not see these minerals explained elsewhere. Please clarify!

Telling the truth, I think that this modelling part with the minerals could be a separate article. Is it really important to include this modelling part in this article? It has a totally different approach and raises a lot of questions. There were enough measurements to prove your primary hypothesis about the reduction effects of amendments on P losses. Especially with the lacking evaluation of the results, this modelling part is not helping, I think. The article should be more concise, more reader friendly, shorter, etc.

Conclusion

It is quite short. Especially compared to the discussion. I think you can conclude in more details!

Environmental issues

I have some concerns about the Al-content of the soils where alum was applied in order to reduce the P runoff (also in water). The main purpose of reducing P losses might end up with Al contamination. What the authors opinion about this?

Other

There are some minor mistypings, extra spaces here and there, I marked some of them, but a thorough check of the manuscript would be nice.

Congratulations on the huge work done.

Please see my comments attached.

Regards, Reviewer X

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thank you for making the changes and taking the comments seriously.

Reviewer 3 Report

Dear Authors,

Thank you for addressing all my concerns.

I am satisfied with the answers and also with the article.

Congratulations again!

Best regards, Reviewer X