Identification and Area Information Extraction of Oat Pasture Based on GEE—A Case Study in the Shandan Racecourse (China)

Round 1

Reviewer 1 Report (Previous Reviewer 3)

This paper needs minor revision. I indicated in the specific comments some small important corrections I think could help to further improve the article.

Comments for author File: Comments.pdf

Author Response

Sincerely thank you for your comments. Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report (Previous Reviewer 2)

Dear authors,

thanks for update. Paper is now much better, but till now Figure 2 is not fully clear fo me. There are missing direction of some lines and also is not clear link between this scheme and text in next chapters. Please try to update image and link it better with next text, to be possible better understand all process. Please make clear all direction in lines in scheme and try to allocate number of chapter to single boxes

Author Response

Sincerely thank you for your comments. Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report (Previous Reviewer 1)

The manuscript is resubmitted, and I recognize that several improvements are made. However, the paper still needs some revision and improvements besides proofreading for a fluent text. I would prefer to maintain Major Changes. Below are some general comments.

Figure1: please add the cartographic elements, improve the caption and label figures as a, b and c;

L236, L217, etc: please present the Figures and Tables in the main text and then show the Figures and Tables;

Figure3: please add title for X-axis;

Figure5: please improve caption detailing what letters a to d means;

L235: the identification of the best features was made visually? please describe better the criteria (just for a given year?). I wonder how you generate the importance ranking score in Table3;

L248: please always refer to spectral bands (not only bands);

L249: numbers smaller than 10, please write down like one, two, three, etc (same for L248, L250, L327, L553 ....);

L257-8: please add one blank line to better see that there is a subsection;

L261: what do you mean with performed well?

L323: better write dates and months than using numbers;

L333: briefly explain the reason of that choice;

L326: better to write "Figures 7 to 9";

L350: please add some previous papers;

L353: after each metric, please add the equation number;

L358: please label the equations and link them with the text;

L367: it is not good to introduce the results section by just showing Table 3. Please bring the sentences starting from L371 and then present Table3;

L369: check typos in your paper (for example, Imporetance);

Tables 5 to 7 could be merged and labeled as A, B and C;

Figure14: please detail better the caption mentioning that subsets are shown for selected region in which small letter show the true color composition for these subsets and small letters is the classified outcome, etc;

L393: all Tables and Figures must start with capital letters;

L410: change aspects to metrics;

Figure12: it would be interesting to merge all three scenarios within one map only or add a fourth one showing the difference (and highlighting increase, decrease and maintenance);

L488 and L491: please avoid showing a long list of references without properly discussing and associating your results with them;

L525-527: in comparison with? the sentence is unclear and I feel you wanna compare with other studies but I am not confident what you want to compare;

L545, L566: in these discussion sections, you are not bringing any mention to your results (Tables and Figures) and only citing one single reference. It needs improvements;

I feel the manuscript still needs a special attention in formulating better some sentences, improving and describing better the caption, improving quality and presentation of Tables and Figures;

Author Response

Sincerely thank you for your comments. Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report (Previous Reviewer 2)

Dear authors,

thanks for update, for me it looks good

KArel

 

Author Response

Sincerely thank you for your comment. Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report (Previous Reviewer 1)

The manuscript "Identification and area information extraction of oat pasture based on GEE——A Case Study in the Shandan Racecourse" is resubmitted to the esteemed Remote Sensing journal. I appreciate some of the changes made and recognize the author's efforts. However, I do not see many improvements in the writing style. I also found that some of my comments/concerns were not considered.

Three days is a bit short to make a profound review and I will highlight some of the comments expecting that authors consider a serious job in improving the manuscript and carefully revising sentences, citations and style.

Below some comments

Figure 1: please add all cartographic elements, including a grid of the coordinates. It is a must according to the Journal guidelines;

Figure 2: please increase the font size for better visualization;

Figure 3: please add a title for Y-axis;

Table 2: no need to use a comma after the author's name;

Figure 6: Cultivated land identification sample points (letters a to e represents show subsets in which samples containing cultivated land, mountain meadow, others, temperate steppe and alpine meadow steppe are found).

L362: First, through the confusion matrix, the overall accuracy (OA) (Equation 1), producer accuracy (PA) (Equation 2), user accuracy (UA) (Equation 3), Kappa coefficient (Equation 4) and F1-score (Equation 5) were obtained. All five equations are futher detailed in Congalton [48].

L416: OA, PA, UA, Kappa, and F1;

L466: 6 to six;

L548: Methods of selecting feature or "Methods of selecting features" or even better, "Feature Selection Methods"?

Please, as mentioned, take the opportunity to ensure careful proofreading of the text and captions of all the tables/figures. I also suggest you pay attention to typos, long sentences, and references. I appreciate that this is not the response you might be anticipating, but my only intention is to ensure this paper is published with good quality.

Author Response

Sincerely thank you for your comment. Please see the attachment.

Author Response File: Author Response.pdf

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

The manuscript "Identification and area information extraction of oat based on GEE——Taking the Shandan Racecourse as an example" is submitted as a full manuscript to the esteemed Remote Sensing journal. The research content is interesting for the Journal but still requires some improvements before being recommended for publication. In addition, I found it quite difficult to link this research's novelty aspect and the outcomes' contribution and proposed methodology to the literature. Therefore, this aspect must be enhanced.

 The written style needs some revision because some typos and sections seem to be written by different authors. I also wonder why authors focus on GEE when the scripts and codes are unavailable and what compromises the "reproducible research" aspect. Additionally, the flowchart presents some information that was not adequately discussed in the methodology and then presented for the first time in the discussion section. Another important aspect is that the captions of both tables and figures are not fully detailed and not correctly mentioned in the main text. Additionally, some of the content, such as the tables and figures, could be presented as complementary material. Finally, the discussion and conclusions could also be improved.

 

Below are some specific comments:

 

The abstract could be summarized and focus more on novelty and contribution to the literature;

The introduction is too long and does not entirely focus on the main issue targeted in the journal;

L94: some of the spectral bands, not all;

L95: which Landsat system?

L184: what can be replicated in other similar environments?

Figure1: please add a country base map;

L216: it would be better to add the subsection names or numbers on the flowchart;

L221: two parts? please identify in the figure two blocks;

L228: the feature selection procedure is not explained;

Figure3: the information represented in this figure, although it looks nice, does not have 3D info, then, please convert it to 2D;

Table 2: add last name of the authors before the reference number;

L243: elevation? slope? aspect? you have to explain how, where and why you use and what 16 means? 

L247: the captions must be detailed and mention sensor and information provided. Y-axis could be labeled as Vegetation Index Value;

L252: add brackets in the labels A to D;

L265: please provide a reference;

Figures 6-9 could be presented as supplementary material (2 figures in one page) and some local pictures could be also presented. Same for others;

L312: presente reference;

L319: it is very unpolite to start the results immediately with a figure. Please add some text and present them adequately;

L325: some sentences could be shifted to methodology;

L375: I am unable to understand the feature selection and how weights and percentage were calculated;

Figure15: please explain in the text; label adequately;

I am still not confident about the data used and how they were used (single date, average, statistics of max ou min?, etc);

L405: was there a second choice? Please recheck the written style;

L406-410: this is not a nice way of discussing results; same for the following sections;

L415: how many small areas exist? It would be better to have a better perception of the landscape and social component in the methodology;

 

L473: there are some discussion sentences inside and the last sentence is not clear;

In general it needs several improvements before being accepted. I would better recommend rejection and encourage new submission once significant improvements are made. This would reduce publication time delays and will allow authors to perform a good work.

Reviewer 2 Report

 

Dear authors, topic of your work is interesting, but presentation of your results in insufficient. You are trying to present all  by images, but textual part is very poor. I add few comments:

·       For Figure 1 I would recommend use maps instead of satellite images. This satellite images is not giving information about are. Or combine satellite image with maps. Probably you can use smaller scale to give better overview.

·       Table 1 is probably not necessary, this facts are known for all community.

·       Generally all part 2.2 is only repeating known facts without clear explanation, which data were used (years, Months, etc.)

·       In chapter 2.3 please add better textual explanations of workflow. The scheme is clear, but will be good to explain all steps.

·       Chapter 2.4 are mainly images, without any textual explanation. This is difficult to read and understand

·       Chapter 2.5 again reduce images add more text

·       Extend also chapters 2.5 and 2.6

·       In Chapter 3 is also necessary reduce images and add more textual analysis

·       Chapter 4 and 5 need to be extended

Reviewer 3 Report

This paper needs major revision.

Comments for author File: Comments.pdf

Reviewer 4 Report

This article is using Google Earth Engine to classify image to five classes cultivated land, mountain meadow, others, temperate steppe and alpine meadow steppe and finally get maps of oat.  The research is interesting and well structured, but there are some concerns that should be addressed:

Figures 4, 5 and 10 should be smaller - there are a bit too big

Line 95 - Sentinel-1 has 10 m spatial resolution, but it is only one with SAR sensor, so please correct that it has lower resolution than Sentinel-2

Lines 301-311 - What were the other chosen parameters for Random forest algorithm? Did you used some feature selection for RF algorithm? If not, why didn't you?

Section Result - Since you have unbalanced dataset it would be of great importance that you show classification accuracy for all classes in 2019, 20202, 2021 so the readers can understand each classification strength and weaknesses for every classification.

    - add tables with classification accuracy for all classes, as well as for all overall accuracies  for easier comparison

Lines 411-417 "Under..." - This is correct, but S1 does not have problem with clouds and there are several researches

(Chakhar A, Hernández-López D, Ballesteros R, Moreno MA. Improving the Accuracy of Multiple Algorithms for Crop Classification by Integrating Sentinel-1 Observations with Sentinel-2 Data. Remote Sensing. 2021; 13(2):243. https://doi.org/10.3390/rs13020243

Blickensdörfer, L., Schwieder, M., Pflugmacher, D., Nendel, C., Erasmi, S., & Hostert, P. (2022). Mapping of crop types and crop sequences with combined time series of Sentinel-1, Sentinel-2 and Landsat 8 data for Germany. Remote sensing of environment, 269, 112831.)

showed that S1+S2, or S1+L8+S2 gives better results. So please rewrite.

 

Lines 468-472 ";perhaps..." - conclusion, future work

 

Reviewer 5 Report

Line 13-21：The introduction at the beginning of the abstract is too long, so it's better to be concise and comprehensive and get to the point directly.

Line 84-85, I don't think the third category can be distinguished from the first two categories. The use of index features is to take this information as the distinguishing features of different ground objects. Finally, machine learning is also used for classification, so it's inappropriate to divide this into the third category. 

Line 180-184 Please go into more detail for a description of the research methodology.

Line 219, I think your flow chart should be revised. Although oat recognition is carried out based on cultivated land recognition, the more connecting lines used between your two processes look chaotic, and your process cannot be seen intuitively.

Line 233-234：Explain what each element in the formula means. (NIR,R,G,B)

Line 252-256 You choose four typical pastures to compare with oats, but there is no explanation and introduction.

Line 252-256 Compared with previous studies, your research content is smaller and the area is smaller, but your farmland recognition accuracy is lower, so where are the advantages of your research?

Line 283-290: Among the five types of sample points mentioned in the article, 356 are cultivated land sample points, while only 72 and 62 are alpine meadow grassland and temperate grassland. Too large difference in the number of sample points may affect the classification results.

Line 291-298: When identifying oats, the number of oat sample points is quite different from other sample points, and the number is a little small.

Line 337-338：For cultivated land, the mean OA is 80% (79%, 81%, 81%) and the mean Kappa coefficient is 0.74 (0.71, 0.75, 0.75), this seems a little low.

Line 318-319：In the result part, the result analysis and accuracy evaluation are divided into two sections, and the accuracy is expressed in the table, which will be clearer, as well as Line 341.

Line 356：Select only four positions for comparison accuracy, is it too little.

Line 397: The discussion part is too thin and not deep enough.

Your research theory, content, etc. are almost the same as the articles you refer to, but the accuracy of the results has not been significantly improved, or even lower, can you elaborate on your research advantages?

 

In summary, I was happy to review your manuscript: “Identification and area information extraction of oat based on GEE——Taking the Shandan Racecourse as an example.” Overall, it is clear that a lot of work has gone into this research. However, based on the GEE cloud platform, the author uses random forest classification to identify cultivated land first, then oats on this basis, and finally carries out accuracy evaluation and feature importance evaluation. This manuscript does not seem to be innovative.