Next Article in Journal
The First Reaction Steps of Lithium-Mediated Ammonia Synthesis: Ab Initio Simulation
Previous Article in Journal
Using the Haney Soil Test to Predict Nitrogen Requirements in Winter Wheat (Triticum aestivum L.)
 
 
Article
Peer-Review Record

Potential Impact of Learning Management Zones for Site-Specific N Fertilisation: A Case Study for Wheat Crops

Nitrogen 2022, 3(2), 387-403; https://doi.org/10.3390/nitrogen3020025
by Camilo Franco 1,*, Nicolás Mejía 2, Søren Marcus Pedersen 3 and René Gislum 4
Reviewer 2:
Nitrogen 2022, 3(2), 387-403; https://doi.org/10.3390/nitrogen3020025
Submission received: 24 May 2022 / Revised: 3 June 2022 / Accepted: 6 June 2022 / Published: 13 June 2022

Round 1

Reviewer 1 Report

Although the comments made have been taken into account, there is still a problem with the methodology and tables 3 and 4. In the methodology it says, and I quote, "Vegetation was measured by YARA sensors which compute, among other information, the Normalized Difference Vegetation Index (NDVI), and the soil electro-conductivity was measured by DualEM instruments, which explore simultaneously different depths of the soil". However, in tables 3 and 4, the values do not have units of measurement, obviously, it is not the NDVI, since the values of this index vary between -1 and 1. What is being measured?

It is necessary to clarify this point both in the methodology and in the results.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors took into account the comments contained in the review.
The article may be published in the currently presented version.

 

Author Response

Dear reviewer,

Thank you for spending time and effort on reviewing your manuscript.

Best regards,

The authors

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.


Round 1

Reviewer 1 Report

All comments of the reviewer:

 

Line

Now is

Should be

Comments

12-13

interpretable machine learning ; precision agriculture ; management zones ; remote sensing ; unsupervised learning

interpretable machine learning; precision agriculture; management zones; remote sensing; unsupervised learning

Delete space

109-111

 

 

Delete lines

127-136

See Table 1 for the summary of the ...

... levels of protein as a measure of quality.

 

Move the selected text fragment together with Table 1 to the beginning of Chapter 3 (Results)

136-137

The methodology for assessing the effects and impact of the fertilization strategies is explained next.

 

Delete lines

162

 …by N∗ = 224kg/ha

…by N∗ = 224 kg/ha

Space

163

…by N∗ = 212kg/ha

…by N∗ = 212 kg/ha

Space

309-312

 

 

Delete lines

318

...the N∗ = 150kg/ha

...the N∗ = 150 kg/ha

Space

425

...is N∗ = 224kg/ha

…is N∗ = 224 kg/ha

Space

426

...212kg/ha.

...212 kg/ha.

Space

445

…of 172kg/ha at Kalundborg and 160kg/ha at Bjerringbro,

...of 172 kg/ha at Kalundborg and 160 kg/ha at Bjerringbro,

Space

I suggest that the Authors of the manuscript add a section "Discussion", discussing the obtained results against the background of other researchers. I also suggest increasing the number of references cited in the manuscript.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper proposes a machine learning methodology for precision agriculture, aiming at learning management zones that allow more efficient and sustainable use of fertilizer. The methodology consists in clustering remote sensing data and estimating the impact of nitrogen management on yield and protein content.

The results obtained in the article were that there was a direct correlation between nitrogen rate and yield and protein content in wheat, and the results of clustering analysis of remote sensing data were directly correlated with nitrogen rate (I didn't see the remote sensing data). However, such clustering analysis results did not establish a relationship between remote sensing data and yield or nitrogen rate, and therefore, there is no way to apply it to precision agriculture or guiding fertilizer application. What confuses me is that to get such a result of correlation between remote sensing data and yield, a correlation analysis can simply be done and a relation between them can also be obtained, so why do we need to use so many methods to do cluster analysis? (I am sorry to say that, or maybe I did not understand the author's intention)

In terms of article writing, the descriptions of each section are tedious and not well organized. It is recommended to simplify the whole article and remove irrelevant descriptions. For example, L314-L320, L337-L344, L346-L356, and L383-L389.

In Table 2, notes are needed for GAP, CH, and SIL.

In Figure 1 and Figure 2, what do the horizontal and vertical axes represent? Need clarification, and the diagrams are not readable. And the figure is hard to read.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This manuscript proposes a machine learning methodology for precision agriculture, aiming at learning management zones that allow a more efficient and sustainable use of fertilizer. The topic is very interesting and has a high impact as we are immersed in a change in terms of sustainable crop production. From my point of view, there are several serious problems. First of all, the article, has more to do with computational methods and statistics than with Nitrogen. Other important issue is that in order to evaluate the effect of nitrogen fertilization on the production and quantity of proteins in the wheat crop, many other variables must be taken into account. For example, precipitation, which contributes nitrogen, is not mentioned; Nitrogen leaching, mineralization, and denitrification are not mentioned. There are many variables that have not been explained or taken into account, so the rest of the work, according to my opinion, has no scientific validity.

In addition, my reading of the paper did raise me a number of concerns, listed below. 

  • The bibliography is sparse and messy
  • Many details are missing in the methodology: size of the plots, how the biomass and protein analyzes have been done, the dates on which the data were taken, etc.
  • NDVI and soil electro-conductivity data do not appear in the results
  • The results could be better presented with a table.
  • There is no discussion, no reference is made to other articles to be able to analyze from other points of view

Thank you for the opportunity to review this manuscript. I hope you fine these comments useful for improving the manuscript.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have revised the specific issues raised in the previous review.

Author Response

Dear Reviewer,

We are happy that you have accepted our changes.

Best regards,

The author team

Reviewer 3 Report

Even having improved the article, from my point of view, there are still several shortcomings. I will list them below:
- Why is the first reference that appears in the text number 12 and not number 1?
- There is a great confusion between the numbers that indicate the bibliographical references and the equations.
- Line 125, add bibliographic reference to the method
- In the methodology, was the entire plot harvested and weighed? or just a sample? Specify
- The results would be presented better if a table were made indicating the significant differences between treatments.
- The results regarding the electroconductivity of the soil and the NDVI still do not appear.
- In the new discussion section, there is no discussion or comparison with other studies (there is only a bibliographical reference). Rather they are results.
I suggest taking these indications into account so that the study can be published

 

 

Author Response

Dear reviewer

We appreciate your effort and time in this review process and we have carefully answered your comments and questions and believe that this has improved our manuscript.

Best regards,

The author team

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

Suggested comments have been taken into account. However, two new problems have arisen.

In Table 2, significant differences are indicated with an asterisk, but the level of significance must be indicated in the header of the table (p<0.01?).

In table 3, there is a big problem, NDVI values ​​vary between -1 and 1, so the data presented cannot be correct.

 

 

Back to TopTop