Simulation of Denitrification Process of Calcium Nitrate Combined with Low Oxygen Aeration Based on Double Logarithm Mode

Round 1

Reviewer 1 Report

I need to recommend a rejection of the article, but I would strongly encourage the authors to resubmit the results they obtained. It is only the interpretation of the results that does not seem appropriate to me.

Firstly, several things are good about this article: The research topic is of strong interest, the study design makes sense (at least from my limited understanding as a modeller) and the results are rather interesting and should be of interest to other researchers in the field. Most of the shortcomings could easily be corrected, e.g.:
- The abstract should already mention the fact that the results were obtained from container experiments and not in situ, and should contain some key information on the study design.
- The introduction should better describe the problem (what did former studies fail to find out) that the study adresses.
- Instead of listing the chemicals that were used for the measurements, for every indicator there should be information on the method that was used, together with a reference paper that describes it. 
- The experiment design should be given in more detail: At what temperatures were the containers stored? Was there air-water gas exchange? What were the light conditions, was there e.g. phytoplankton growth? Was the sediment stratification preserved while transferring it to the container or was the sedimentary material mixed?
- Finally, the article would need some proofreading, e.g. I couldn't find a second reference for "sediment loneliness".

But the major shortcoming and the reason why I recommend a rejection of the article is the lack of motivation for the double logarithmic method. The article states that the best treatment method is the one with the lowest value of "a", but it does not prove why this should be the case. I will give a few more thoughts on that to illustrate the problem.

If I am not mistaking, the best treatment method is the one that, in the long term, shows the largest reduction of ammonium concentrations without raising nitrate too much. So, trivially, the concentrations showing up in the last sample (being taken after 30 days) would be used. Now I understand that there may be issues with this (e.g. measurement errors in a single measurement, system has not yet reached steady state...) which is why alternative methods are of interest. But possibly it is not this double-logarithmic one.

Firstly, there is no mechanistic reason that the concentrations are expected to behave like c=k*t^a. This would, e.g., mean that the final concentration would always be either zero (if a<0) or infinity (if a>0).

Secondly, there is no empirical proof from this study that the concentration always follows this equation. While for some of the experiments the dots look like on a straight line, for others they don't. 

Thirdly, it makes no sense intuitively from looking at examples. Why would the shape of the decay be of interest? Why would a treatment where the concentration decays like 1/t^3 be preferred over one that shows a decay of 1/t^2 or 1/t? (corresponding to a=-3, a=-2, a=-1)

Finally, the initial concentration at t=0 cannot be taken into account if t is put on a logarithmic scale. So, if an initial increase of c happens between t=0 and t=1, this is ignored. Everything is only measured relative to the value at t=1. This obviously cannot be optimal, since of course in reality we want to see a reduction against the concentration before the treatment, c(t=0).

Maybe I am overlooking something and the log-log method is actually good. "Good" means it is better than other methods in projecting what the long-term concentration change might be, based on the limited information we have from our data (especially the limited duration of the experiment). But then the authors should argue why this is the case, and convince the reader (and the future reviewers).

With best regards,

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Article entitled “Quantitative Study on Denitrification Process of Calcium Nitrate Combined with Low Oxygen Aeration Based on Double Logarithm Mode” is written on quite good English level. However, I still recommend careful grammar and language check. Some of the sentences (e.g., lines 46-51 and 94-98) are too long, which makes difficult for the reader to fully understand the text.

Moreover, below please find my detailed comments and questions.

• Lines 62-63: “The pollution sources of black-odorous water body are mainly divided into exogenous and endogenous pollution”. The same statement was already mentioned in the lines 42-43.
• Lines 63-65: Could Authors add more specific examples of exogenous and endogenous pollutions?
• Phrase in-situ is ones written with a dash, ones without. Please unify.
• Line 78: I am not sure if the term “electron receptors” is correct. It should rather say “electron acceptor”
• All the shortcuts as “SRB”, “TN”, “DO”, “ORP”, “DNRA” should be explained when first appear in the text.
• In the 2.1. chapter: “Background concentrations of major pollutants in the overlaying water and sediment were measured before the experiment”. I do not think pH could be classified as “pollutant”. Presented in Table 1 background values were measured after collecting the samples or in the flowing river? How was the sample of sediment collected? Why such measurements were not repeated after the calcium nitrate treatment? Authors should also write information about purity and producer of used chemicals as well as brief procedure of each indicator measurement.
• In the 2.2. chapter: “Calcium nitrate particles” should be rather described as a “calcium nitrate dosage”.
• Lines 134-136: How much of the sample was collected? From which place/depth? How Authors indicated the changes of ammonia nitrogen and nitrate nitrogen (please add a more detailed procedure to the subchapter 2.2.3.)?
• Could Authors present the chemical reaction/dependence of calcium nitrate, ammonium nitrogen and nitrate nitrogen?
• Chapter 4 is more like a continuation of results, not a Discussion. There is also lack of comparison to the literature in this part.
• In the reference 5, there should be given the date of access. Also, the reference 8 do not include the volume number and the pages.

Concluding, I suggest major revision of presented article.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The paper was clearly improved by the additions of the authors, I now recommend major revisions because my main point was not resolved, and this main point is that there is no argumentation for the double logarithmic method. Why is it the slope in the log-log plot which defines which treatment is optimal? 

I will make an example: After treatment A, the concentration drops like c_1 = a*t^-2. After treatment B, it drops like c_2 = b*t^-3. Obviously this decay law cannot be valid from t=0 already because there would be an infinite concentration then, but we assume it is valid from some time afterwards, so from then on the decay follows a straight line in a log-log plot. 

The slope in the log-log plot is -2 and -3, respectively, so the second method is always favoured, irrespective of the values of a and b. But in reality the values of a and b do matter. What we might want is e.g. a treatment where the concentration drops below a defined threshold value C as fast as possible, or where at a certain time T after the treatment, e.g. after 1 year, we want to have a concentration that is as low as possible. For both of these problems, it would not make sense to look at the slope alone but to consider the values of a and b as well. If a is small enough and b is large enough, actually treatment A might be the better one. 

I see the point that for very large times T, it is always the slope that matters more compared to the a or b value. But then the question is, is it valid to extrapolate this rate law observed in the first, say, 30 days, to the time T? Or, will the slope in the log-log plot actually change over time, such that the third-power decay observed first will e.g. become a first-power decay later?

I want to see a discussion of this point in the manuscript before I can recommend publication. Especially what the authors need to show is why the log-log method is better in predicting the long-term outcome, compared to the obvious way of choosing the treatment with the lowest concentration at the end of the experiment. This is critical because if this was not the case, if the slope in the log-log plot is actually less predictive of the final long-term result, using the double logarithmic method might actually lead to selecting less effective treatment options in the long view.

Author Response

Dear reviewer,

Thank you for your kindly comments and suggestions. All of the authors feel very lucky for helping us improve our paper. The following are our replies to your comments and suggestions.

The reason why we used the double logarithm method was that it may simplify the nonlinear problem into a linear problem. As you say, the values of a and b really matter. To be honest, in our research, as you see in Tables 1 to 4, the values of b in the regression equation lnc_t=alnt+b of ammonia nitrogen were almost the same, which means that when t=1, the values of c_t were basically equal. That is to say, in this circumstance, it was the value of a that really mattered. That is why we mainly looked at the slope alone without considering the values of b much in the discussion of ammonia nitrogen.

As to the nitrate nitrogen, the addition of calcium nitrate or the low oxygen aeration is a main factor that affect the value of b in the equation lnc_t=alnt+b, so the value of b differred under different dosages of calcium nitrate addition or under different aeration rates. However, we still think the slope is more important than the value of b due to the following reason. We only used the slope to reflect the change rate of nitrate nitrogen. As to finding the optimum experimental conditions, actually now we can only use the obvious way of choosing the treatment with the lowest concentration at the end of the experiment.

Yes, as you say, the rate law we obtained in the 1st to 30th days may hardly continue to apply in the long run. Actually the ammonia nitrogen/nitrate nitrogen didn’t change much after 30 days due to the stabilization of the reaction. Since the rate law after 30 days still needs to be explored, we have added the scope of application of the double logarithmic method. Besides, we want to clarify that the log-log plot was only used to characterize the change rate of ammonia nitrogen/nitrate nitrogen in overlying water/sediment in a certain range, not to predict the final result in the long view.

Thank you again for giving us so much kindly comments and suggestions. Please do not hesitate to contact us if you still has some questions.

Reviewer 2 Report

Authors answered all my questions and doubts. In this regard, I suggest accepting presented revised version of manuscript. 

Author Response

Dear reviewer,

Thank you for your kindly suggestions. All the authors give you great thanks for improving our paper.

Best regards.