Real-Time Quantification of Crude Protein and Neutral Detergent Fibre in Pastures under Montado Ecosystem Using the Portable NIR Spectrometer

Round 1

Reviewer 1 Report

The work entitled “Real-time quantification of crude protein and neutral detergent fibre in pastures under montado ecosystem using the portable NIR spectrometer” is a very interesting work that highlights a technique like NIR, which is fast, easy to use and non-destructive. In addition, the technique allows the farmer to stay informed about animal supplementation and make decisions in real time. Unfortunately, in general, the work is very descriptive and has samples only from spring. A more in-depth statistical analysis and greater clarity in the material and methods section are lacking. Although the discussion is very long, the results are briefly explained. Therefore, although the work is very interesting from a technical and applicability point of view, I recommend that the authors analyze these actions and correct these aspects in depth (major revision) before publication. In addition, I recommend correcting the following suggestions:

 

Introduction:

In general, the introduction is extensive and well referenced, as well as giving information about the objective of the work. However, the section that introduces the NIR technique is very repetitive, which extends the introduction without this being necessary. I suggest summarizing and simplifying this section of the introduction.

In addition:

• Line 40: missing a space [3].It
• Line 92 and 98: the authors should unify the word nondestructive or non-destructive
• Line 105: I suggest that the authors introduce the abbreviations CP and NDF (neutral detergent fiber, NDF and crude protein, CP) in the introduction the first time they are named.

Methods:

In general, the methodology is a bit confusing, the authors mention results and the different subsections are not very well differentiated.

• Line 157: there are two space….. and “Tapada
• Figure 1 is not of good quality. Also in the legend the situation of the farms is mentioned (approximately). I suggest eliminating since I assume that the authors have the coordinates for the elaboration of the map.
• Again double space on line 173.
• Paragraph 161-176 should be in a section in the results since meteorological data is mentioned. Same with paragraph 182-190.
• Throughout the text the authors talk about our study. I suggest the use of "this study"
• Line 197: what the authors mean by: Note that on the Mitra farm, there are 3 fields (Mitra A, Mitra B and Mitra C).
• line 201, CP and NDF determination should be another subheading. Furthermore, it must be explained with the same precision as the rest of the sections of the methodology.
• line 244: correct: pre processing.
• A descriptive analysis is not sufficient to be considered a statistical analysis. Have the authors tried to differentiate zones, or to analyze the downloaded meteorological variables and compare them with the determined ones? what statistical software have they used.

Results:

• The authors must include in this section the meteorological conditions described in the material and methods section.
• Table 1: Could you put the total number of samples for each mean value of each farm?
• Lines 274-276: could be unified in a single sentence.
• Line 276: could the mean value be added to the table 1?
• This paragraph lacks a comparison between the contents in the different farms. This is where a comparative statistical differential would be expected.

Comments for author File: Comments.pdf

Author Response

Dear reviewers,

I would like to begin by thanking for your suggestions to the manuscript entitled “Real-time quantification of crude protein and neutral detergent fibre in pastures under montado ecosystem using the portable NIR spectrometer” that was submitted to the Special Issue of Applied Sciences journal in the special issue entitled “Application of Spectroscopy as a Sustainable Tool for Agri-Food Product Inspection and Characterization”. Your suggestions were valuable contributions to our manuscript. We have incorporated most of the suggestions and those changes are highlighted in the document.

Response to reviewer 1,

In the introduction we deleted some paragraphs (red letter in the document that I am sending) referring to the NIR technique, as suggested by you. In figure 1, we put the coordinates of each plot next to each one. The weather data and related graphics have been moved to the results section. For CP and NDF determinations a new section was created. The statistical analysis section has also been redesigned, to answer and clarify your questions and doubts. The text that has been added is highlighted in yellow. In table 1, a column with the number of samples per plot was also added. All other suggestions and corrections were accepted.

All changes are marked in the document in word and pdf. The text and data that have been added are in yellow and what is to be deleted is in red.

 

Thank you very much for your attention to our paper.

Sincerely,

Emanuel Carreira, Ph.D student

 

MED-Mediterranean Institute for Agriculture, Environment and Development Rural Engineering Department - University of Évora P.O. Box 94, 7002-554 ÉVORA, PORTUGAL telephone +351266760800; fax +351266760811 e-mail: emanuel.ruben@hotmail.com

 

 

Author Response File: Author Response.docx

Reviewer 2 Report

Overall, I think the manuscript is very well done, and provides useful results on the use of in-field NIR spectroscopy for forage analysis. However, I would recommend one experimental addition that would greatly strengthen the manuscript. It seems to me that there are two issues that could negatively affect the final model performance statistics.

1. Lack of appropriate wavelengths (e.g., N-C bands for the CP model)
2. Interference from water on relevant wavelengths

The authors mention both issues, and I completely agree that the poorer performance of the CP model compared to the NDF model is likely due to #1. However, it should be possible to estimate the relative importance of the water interference by re-scanning the 87 dried and milled samples and see if the PLS models improve substantially. This would quantify the ‘trade-off’ between accuracy and convenience for in-field sampling.

Three other technical issues need to be addressed –

1. What do the author’s mean in line 185 “when there is still some soil moisture content”? I think they mean something like “when the soil moisture content is above xx%”. This should be stated and referenced.
2. What was the significance of the 0% reference value (lines 229-231)? How were these scans used in the calculation of reflectance?
3. How were outliers excluded by PLS (line 248)? The method (and the # of excluded samples) is needs to be clearly stated.

 

Finally, there were a few minor issues that should be addressed -

A distance scale should be included in Figure 1 – how big an area is the Montado?

In Figure 2 – highlight the months where samples were taken in Figure 2a and 2b, add average rainfall amount in 2b, since the authors say that the spring of 2021 was particularly dry (lines 173-174).

English grammar – it may be of value for the authors to consult an English editor. There were several awkwardly-phrased sentences and incorrect word choices in the ms, e.g., line 54 – “floristic”, line 121 – “manufacturing process” should be “ease of manufacturing” or “cost to manufacture”, line 398 – “developed” should be “discussed”.

Author Response

Dear reviewers,

I would like to begin by thanking for your suggestions to the manuscript entitled “Real-time quantification of crude protein and neutral detergent fibre in pastures under montado ecosystem using the portable NIR spectrometer” that was submitted to the Special Issue of Applied Sciences journal in the special issue entitled “Application of Spectroscopy as a Sustainable Tool for Agri-Food Product Inspection and Characterization”. Your suggestions were valuable contributions to our manuscript. We have incorporated most of the suggestions and those changes are highlighted in the document.

Response to reviewer 2,

The first comments raised by you: we considered that, in the CP model the important spectral regions were selected according to the weighted regression coefficients plot. However this spectrometer (MicroNIR) has a small  spectral work range and the main reason to have poor calibration models is the absence of some wavelengths that are very important to calibrate  the crude protein. The water is a problem, yes, however it is very important for us to have a method to quantify the crude protein directly in the field in real natural conditions. We have included this coment in the manuscript.

As for the 3 technical questions you posed, regarding soil moisture, the significance of the 0% reference value and the way in which outliers were excluded, the comments are included in the document, highlighted in yellow. The precipitation for the months of March and April 2021 was included in the document. As for the area of the montado, a bibliographical reference (53) was added where the area of each plot is described.

 

All changes are highlighted in the document in word and pdf. The text and data that have been added are in yellow and what is to be deleted is in red.

 

Thank you very much for your attention to our paper.

Sincerely,

Emanuel Carreira, Ph.D student

 

MED-Mediterranean Institute for Agriculture, Environment and Development Rural Engineering Department - University of Évora P.O. Box 94, 7002-554 ÉVORA, PORTUGAL telephone +351266760800; fax +351266760811 e-mail: emanuel.ruben@hotmail.com

 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The transportable NIR tool is a really fast and useful technique to determine, predict and report on different compositions in different products. The present study demonstrates this utility, therefore it represents an advance in this class of methodologies that provides a practical advance both in industry and in research. For this reason, I consider that the findings should be published.
However, I suggest authors to be careful with their presentation.
Thus in table 1 the format is neglected. I suggest changing No. of samples to Samples (N) in such a way that it cuts the size.
Also on line 260 there is a bold sentence that should not be bold.
Figure 3 in section a, the background of the letter a is cut off with the limits of the figure. move it or remove the white background.

Otherwise, the authors have answered and corrected my suggestions and questions.

 

Author Response

Dear reviewers,

I would like to begin by thanking for your suggestions to the manuscript entitled “Real-time quantification of crude protein and neutral detergent fibre in pastures under montado ecosystem using the portable NIR spectrometer” that was submitted to the Special Issue of Applied Sciences journal in the special issue entitled “Application of Spectroscopy as a Sustainable Tool for Agri-Food Product Inspection and Characterization”. Your suggestions were valuable contributions to our manuscript. We have incorporated most of the suggestions and those changes are highlighted in the document.

Response to reviewer 1,

Table 1 has been improved as suggested by you. However, on line 260 we didn't see anything in bold. Figure 3, section a, has also been moved so as not to overlap the limits of the figure.

All changes are marked in the document in word and pdf. The text and data that have been added are in yellow and what is to be deleted is in red.

 

Thank you very much for your attention to our paper.

Sincerely,

Emanuel Carreira, Ph.D student

 

MED-Mediterranean Institute for Agriculture, Environment and Development Rural Engineering Department - University of Évora P.O. Box 94, 7002-554 ÉVORA, PORTUGAL telephone +351266760800; fax +351266760811 e-mail: emanuel.ruben@hotmail.com

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have addressed my comments on the 1st review. I believe that additional work (perhaps in another article) to repeat the modeling work on the dried and milled samples would be valuable - I recognize that in-field prediction is the ultimate goal, but avoiding the traditional NDF and CP analysis methods using a laboratory-based NIR method would still be valuable.

Author Response

Dear reviewers,

I would like to begin by thanking for your suggestions to the manuscript entitled “Real-time quantification of crude protein and neutral detergent fibre in pastures under montado ecosystem using the portable NIR spectrometer” that was submitted to the Special Issue of Applied Sciences journal in the special issue entitled “Application of Spectroscopy as a Sustainable Tool for Agri-Food Product Inspection and Characterization”. Your suggestions were valuable contributions to our manuscript. We have incorporated most of the suggestions and those changes are highlighted in the document.

Response to reviewer 2,

We appreciate your comments on our manuscript.

All changes made previously are highlighted in the document in word and pdf. The text and data that have been added are in yellow and what is to be deleted is in red.

 

Thank you very much for your attention to our paper.

Sincerely,

Emanuel Carreira, Ph.D student

 

MED-Mediterranean Institute for Agriculture, Environment and Development Rural Engineering Department - University of Évora P.O. Box 94, 7002-554 ÉVORA, PORTUGAL telephone +351266760800; fax +351266760811 e-mail: emanuel.ruben@hotmail.com

Author Response File: Author Response.docx