Production–Living–Ecological Spatial Function Identification and Pattern Analysis Based on Multi-Source Geographic Data and Machine Learning

Round 1

Reviewer 1 Report

1. On the whole, the logic of the first paragraph of the introduction is somewhat confused, which is not closely related to the topic of the article and does not highlight the necessity of the research. I suggest that the author should make improvements.

2. In line 32-35, the so-called "interaction" should be the relationship between PLES. It not only refers to the production space occupying living space and ecological space mentioned here, but also includes: the impact of living space on production space and ecological space, for example, the large occupation of urban residential land will lead to the reduction of urban green space and farmland area; The impact of ecological space on production space and living space, such as disturbance and destruction of ecosystem may lead to frequent occurrence of natural disasters, and bring threats to production activities and residents' lives. I suggest that the author improve on this and fully explain the relationship between the three.

3. In line 36-39, I think the sudden mention of "land use science" here is a bit abrupt, and this sentence does not have a strong logical connection with the text. What is mentioned above is "sustainable land use", and I suggest that the author change this part to the relationship between sustainable land use and PLES.

4. In line 39-41, I think it is not clear why the research object is urban agglomeration, whether urban agglomeration are harmed by the conflicts among the PLES, and what is the necessity of studying urban agglomeration. I suggest the author make improvements.

5. In lines 60-69, the author only describes the advantages of the "OBIA". I suggest that the author also elaborate on its disadvantages, because the advantages and disadvantages of the other two methods are described in the above.

6. Lines 91-103, this paragraph mainly introduces PLES, but I think this paragraph is a bit long, because the first paragraph of the introduction has already mentioned the current conflicts among the PLES, and it seems to be repeated here. I suggest that the author abbreviate this section and combine it with the next paragraph.

7. In line 96-100, the problem elaborated here looks like "the impact of urbanization on the environment", which has little to do with the logic of "Conflicts among the PLES" mentioned in the next sentence. I suggest that the author directly describe the typical problems of PLES conflict here.

8. In lines 100-102, which only dwell on the challenges posed by the conflicts among the PLES, I think it is possible to extend the content and point out the necessity of studying PLES from the perspective of function.

9. In line 107-114, I think these two examples have little to do with scale, and it is unreasonable to generalize with urban scale and regional scale. I suggest that the author summarize according to the methods used in the two examples in order to follow the content of the paper. For example, Li uses the index system measurement method and Liu uses the merged classification method.

10. In lines 130-131, the research scale and the research object are different. I think the scale of this paper is not urban agglomeration, but 10m grid scale. I suggest that the author think it over again.

11. Lines 163-167, which is a brief description of the natural and socio-economic conditions of the BTH, should be placed at the beginning of this paragraph as a brief description of BTH, not at the end. I suggest that the author move this paragraph to the front and replace the original content here with the land use situation in BTH area, which will make the context more coherent.

12. In Figure 1, the two maps do not use the same coordinate system, and I suggest that the author modify it and add the boundaries of each province and municipality.

13. In Table 1, I suggest that the author center the text "POIs density" in the last line, which will be more beautiful.

14. In line 210-212, I think the innovation of this paper can be explained here, that is, the method used in this study can specifically distinguish the categories within buildings, otherwise it can not be distinguished from the ordinary research on the function identification of the PLES.

15. Figure 2 does not well show the participation of POI Data and night light data in the subsequent OBLA processin "Data processing". I suggest that the author add some arrows for improvement. In addition, the left border line of the entire image is too close to the content, can you move it to the left?

16. In lines 222-231, this part of the content has been elaborated in the introduction, and there will be some repetition if it is repeated here. It is suggested to abbreviate it here, and some supplementary content can be moved to the introduction for detailed elaboration.

17. There seems to be a repetition in lines 233-237.

18. In line 283, I think it is necessary to explain which years' data are the samples involved in model training and which data are obtained later.

19. The font in Figure 3 seems a little small, and the overall diagram is not very compact, which I suggest the author improve.

20. The information in Figure 4 is vague and sometimes hard to see. Secondly, the whole picture is a little wide, I suggest the author adjust the layout of the 5 pictures.

21. In Figure 5, firstly, why are the objects selected in the minimap Shijiazhuang, Beijing and Cangzhou, rather than Tianjin, Beijing and Shijiazhuang? After all, Beijing, Tianjin and Hebei are mentioned in the introduction of the research area . I hope the author can give an explanation or make improvements. Second, the author does not seem to indicate in the article the spatial distribution of PLES in what year the graph is, and I suggest that the author should do so. Third, I suggest that the author add a north pointer and scale to each diagram.

22. Figure 6 shows the distribution patterns of artificial PLES areas in different cities. First of all, the author does not explain which land types belong to artificially utilized land, nor does he explain which type of PLES belong to artificially utilized land. I suggest that the author make improvements here. Secondly, I think the map also needs to show the proportion of non-urban ecological space and transportation space, because some of these two types of PLES also belong to artificially utilized land.

23. In lines 460-465, the disadvantages of the pixel-based classification methods and the scene-based land use classification are only described here. Later, it is directly indicated that the OBIA is used in this study, which I think is not logical. I suggest that the author follow the disadvantages of both methods with a brief explanation of why the OBIA is used and the advantages of the OBIA.

24. I think it is necessary to explain in the discussion section that the PLES type of each pixel obtained in this paper is only its dominant type, because each pixel is a multifunctional synthesis of PLES.

Need to be improved

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript highlights the originality of mixing several sources for creating PLES maps. However, there are some following points that the authors should be considered:

- The introduction does not properly demonstrate the need for machine learning to be implemented for city clusters, and what is the aim of classifying PLES into urban groups other than one city?  The introduction should state the purpose of the same research.

* Method section:

- Images should be crisper.

- Please clarify why images from 2013 to 2021 were used when Sentinel 2A were launched in 2015.

- Can the author explain why 16 texture features of the primary component bands were obtained?

- The authors should add horizontal lines to Table 1 to clearly identify various groups.

- Figure 2 is quite blurry; the link between POIs and Night light data is not visible. Is this information used for segmentation?

- The authors need to explain more clearly how objects can extract spectral features shape, texture, and socioeconomic features of each pixel. The confusing point is why after using segmentation, then extracting it back to pixels? Is it taking training samples for Random forest from objects? Need to explain further how to identify groups of features.

- Section 2.3.1: The supplied content reiterates the ideas from the introduction, although it is unclear whether the authors use the level 2 elements in table 2 to establish PLES or vice versa. Otherwise, the PLE classification pattern must be provided in greater detail.

+ Section 2.3.2 should be included in the Introduction. This part should clearly provide the Random Forest input data, some significant RF parameters used in this study, and define which output land cover or land-use groups?

* Results section.

- The definition of classification for Plans (A, B, C) needs to be clearer, the input factors for the classification of each plan need to be clarified rather than mixed with the validation part.

- Session 3.1 should be improved, written separately into paragraphs would be nice. Does Figure 3 repeat the content of plan C in Table 3?

- Lines 337 – 347: It is hard to understand.

- Figure 4 is too blurry for readers to understand the authors' discussion. Many factors have not been presented in the method.

- The captions of Figures 3 and 5 are not clear (At what time are the maps?) while the caption of Figure 4 is too detailed.

- Discussion must come before Conclusion.

- Figure 1 depicts political factors as well as national sovereignty. I am only concerned in academic aspects. Accepting the article does not imply agreeing with the writers' annotations on the map.

I think the authors need to do a major editing before this manuscript can be reviewed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The article is relatively well described in terms of theoretical background, data used in the study and its results. The whole is interesting and well commented. Some remarks may arise: (a) the use of personal mode, see the use of the word “we”, e.g. in line 21, 216, 330, 339, 34, 437 - these sentences should be transformed by using the impersonal mode, (b) figure 3 needs improvement - dark font of percentage shares on a dark background is not visible when printed, (c) figure 6 needs improvement - percentage shares are too small when printed; one could consider enlarging these figures, (d) introducing conclusion before discussion is interesting and should be maintained, but the discussion itself requires a definite improvement. The reservations to this part are quite significant, since the authors write that there are many studies on similar classifications, this would require at least introducing a few names into the content. Both the achievements of the authors and the repetitions of what they managed to confirm from other works would require writing appropriate references. This part of the work definitely needs improvement.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

All the questions have been resolved. The paper can be considered for publication

All the questions have been resolved. The paper can be considered for publication

Author Response

        Thank you very much for your valuable comments. We have again refined parts of this manuscript in this round of revisions, The changes to our manuscript have been highlighted within the document in red color. 

Reviewer 2 Report

Dear the authors,

Thank you for your efforts in revising the manuscript. I have no more comments.

Best regards.

Author Response

Dear the reviewer,

       Thank you very much for your valuable comments. We have again refined parts of this manuscript in this round of revisions, The changes to our manuscript have been highlighted within the document in red color. have been highlighted within the document in red color. Thank you again for your approval and help with this arcticle.

Best regards.

Reviewer 3 Report

Thank you for the corrections and changes made. I do not see the need for any further ones. I would only consider adding a few footnotes to the sentence starting in line 490 - since the words ‘Many scientific research’ were used, it begs to add references here… it would also enrich this part with additional footnotes, which are not too many here.

Author Response

Please see the attachment.

Author Response File: Author Response.docx