Retrieval of Three-Dimensional Green Volume in Urban Green Space from Multi-Source Remote Sensing Data

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study developed a straightforward and effective approach for three-dimensional green volume (3DGV) retrieval and mapped 10-m resolution 3DGV in urban green space (UGS). The results are interesting and convincing and the structure is clear enough to follow. However, a minor revision is needed to improve the quality of the manuscript. Many specific comments are listed below.

1.       Line 39, the citation is needed for the mentioned researcher’s name.

2.       Line 47, “indictor” should be revised to “indicator”.

3.       Line 59, I think “UAV” should be replace with “UAV technology”.

4.       Line 61, “the overall relative error” should be replaced to “an overall relative error”.

5.       Line 69, “various technical approaches have applied in different situation” should be deleted.

6.       Line 95, “Details” should be replaced with “Detailed workflow”.

7.       Line 145, author should rewrite the steps of preprocessing to make the manuscript more concise, I think “median image” and “resample” can be summarized as preprocessing with “atmospheric correction” and “cloud masking”.

8.       I suggest to revised the title of Section 2.2.3, this section contains the pre-processing of UAV images, not just the acquisition of images.

9.       Line 175, “orthro” should be revised to “ortho”.

10.   Line 228, “The details of formulas were described in Table 1” replace with “The empirical formulas of 3DGV in previous studies are listed in Table 1”, please check the similar error of verbs.

11.   What is the mean of “Since 3DGV is a three-dimensional indicator that represents spatial volume of crown, and volume is typically calculated by multiplying area-based parameters and height”? Section 2.5 should be rewritten to provide a more detailed description of the modeling process, please express more clearly for how to construct the retrieval models, especially the compound model.

12.   In Section 2.5, when you introduce the modeling strategies, the referenced paper should be cited.

13.   In Section 2.6, please explain the “several ranges” or cite other research papers to demonstrate the feasibility of this method.

14.   In Table 3, I suggest to change “LAI models” and “CHM models” to “3DGV models based on LAI”, and “3DGV models based on CHM”, please check the similar names throughout the whole manuscript.

15.   Line 302, replace “analysis” to “analyze”, please check similar grammar mistakes throughout the whole manuscript.

16.   Line 319, “the correlation analysis results” should be expressed more clearly, I suggest to replace with “the correlation analysis results of 3DGV, LAI and CHM”. Please check similar expression mistakes throughout the whole manuscript.

17.   Line 323, the consistent of spatial distribution cannot prove the correlation relationship between 3DGV and LAI, please rewrite this sentence.

18.   The expression of equation (8) to (12) in Line 308, please explain what the parameters represent.

19.   The font size of the equations in the Table 3 and Table 4 should be consistent.

20.   Line 383 and 385, I suggest to replace the “Figure 12 a, e” with “Figure 12 a and e” or “Figure 12 a to e”.

21.   Line 411, I think “Figure 12 f” should be revised to “Figure 11”.

22.   Line 417, I consider that an accuracy of 75.15% cannot be described as “excellent”, please replace it.

23.   Line 432, I think “expect” should be revised to “except for”.

24.   In Figure 13, I suggest to rename the title and please label the image resolution in the legends.

25.   Line 490, “to enhancing” should be revised to “to enhance”, please check similar grammar mistakes throughout the whole manuscript.

26.   I think the title number of “Estimation method and predictor variables selection” is “4.2”.

27.   In Section 4.3, I suggest you to change the title of “limitations and future directions” to “limitations and strengths”, the strengths for choosing parametric models should be mentioned in this section, not in Section 4.2.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript estimated the three-dimension green volume in urban green space from parametric estimation model based on multi-source remote sensing data. The results showed the three-dimension green volume was consistent with the reference three-dimension green volume monitoring by unmanned aerial vehicle, and the optimal model achieved an excellent accuracy in 3DGV mapping. However, some key fundamental issues should be addressed or discussed because these issues are crucial to the interpretation to their results. I would like to see a revision of the paper before possible acceptance of Remote Sensing by addressing the following issues:

 

Major comments:

1.     Generally, the overall accuracy is great than Kappa coefficient, in Line 159, Kappa coefficient (0.91) was great than overall accuracy (0.9064), the same problem was found in Line 181, please check this value to ensure the numbers are correct.

2.     In this study, it is found that the overestimation occurs in areas with sparse vegetation, while the typically happens in areas with dense vegetation, the authors should give more discuss about the reason in the manuscript.

3.     The predictor variables have import influence on the accuracy of the 3DGV. In this study, the LAI was selected as the variables in the estimation model, we found LAI was calculated based on NDVI from the satellite images. As we know, the NDVI has a trait, it tends to be saturation in dense vegetation area. Maybe, it will also influence the LAI value in this area, that is why the estimated 3DGV was underestimated in areas with dense vegetation.

4.     As the authors pointed out that the satellite-derived 3DGV and UAV-derived 3DGV have significant difference due to the different spatial resolution. We suggest the authors add more comparation of the influence of resolution on the 3DGV results.

5.     The author indicated that the 3DGV estimation model was aimed to extend to larger scales, but in this study, the research was only focus in a very small area in Kunming, Yunnan province. If the model used for larger area estimation, we suggest the author do more experiment in different season, different city in various condition, such as in different area of China, different development level. And, add additional discuss about this part.

 

 

Comments on the Quality of English Language

 The manuscript should be polished or corrected by a native speaker before published.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

I congratulate the authors on a very interesting article, which was also well prepared in the editing aspect.

I only have a comment regarding the methodology of the proposed Study sites.

The test area is very small - approximately 2 x 0.5 km. This means that the satellite image was about 200 x 50 pixels (10m resolution). The experiment should be performed on the entire satellite scene (perhaps cropped to the city limits). Small areas of reference data should verify the results.

 The reference data were selected well (UAV data and field verification).

I also missed the presentation of selected Study sites on the background of the whole city - the whole satellite scene.

The climate in the Study site has not been sufficiently described. Why were satellite images from April to May was chosen?

Best Regards.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

There is a major problem with this paper and it concerns the explanation of the UAV and satellite data. The Wuhua case study is 1.12 km2 and the Jinning case study is 1.45 km2. High resolution images of these two sites are given in figure 1. But this translates to about 9 Sentinel pixels (at 10m pixel size) for Wuhua and about 12 pixels for Jinning. Eventually the paper shows this to some extent in figure 13. There is no point in doing this on such a small area. It cannot show anything of general value when the overall objective is to use satellite data, which implies larger areas. If we then consider that the paper uses Sentinel-1 radar data then the quality of the radar data is not sufficient to try and analyse 9 or 12 pixels.

There is either no explanation of what has been done with the satellite data or the paper is seriously flawed in its design.

Overall, the authors have used a large number of variables and attempted to produce correlations and regressions amongst the long list of variables. This does not illuminate the problem. The paper is far too long. If the paper is about using satellite data then it cannot restrict itself to these two small areas. If it is not about using satellite data then it should omit any satellite data references and be rewritten with primarily UAV data involved.

Comments on the Quality of English Language

Needs some improvement, for example in figure 1 it is "reference" not "referenced" and in line 67 it is not "proved" but "showed".

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

Comments and Suggestions for Authors

Hong et al present the ms 'Retrieval of three-dimensional green volume in urban green space from multi-source remote sensing data'.  This study focuses on quantifying the three-dimensional green volume (3DGV) in urban green spaces (UGS) at a regional level. Estimating 3DGV accurately from satellite images has been a challenge. The researchers developed a model to estimate 3DGV in UGS by using Sentinel-1 and Sentinel-2 images. They initially used UAV images to calculate a reference 3DGV using a pixel-neighbor averaging algorithm. Then, they utilized the Canopy Height Model (CHM) and Leaf Area Index (LAI) derived from Sentinel-1 and Sentinel-2 data to create estimation models for 3DGV. The study compared the accuracy of these models to select the best one. The optimal model combined a LAI power model and a CHM linear model, yielding a good fit (R2=0.78, MPE=8.71%). The estimated 3DGV closely matched the reference 3DGV in spatial distribution, achieving an excellent accuracy in 3DGV mapping (Overall Accuracy = 75.15%). This research highlights the potential of using Sentinel-1 and Sentinel-2 images to develop a reliable model for estimating 3DGV in UGS.

Introduction is well written and follows all the required criteria: it introduces us to the subject ms, mentions previous articles well and the objectives well defined.

Materials and Methods are fine. 

The authors still need to state their methods and that's it. Mentions of previous studies should be made in the Discussion section.(201-202, 235-236)

All methods described in this section must relate strictly to this study so the phrase 'in this study' is redundant.

Then I had the feeling that we had already moved on to the next section, Results. All the graphs that appear should be moved to the next section. Do not change the purpose of this section.

Results

There are figures (figure 8, 9, 10), table (3) and text for the same results. There is a lot.

Same for the next paragraph (351-368), figure 11 and table 4.

The discussion is very brief and with few bibliographical references mentioned.

In addition the paragraph 'Limitations and future research directions' which is extremely welcome should be moved to Conclusions.

The Conclusions section should not repeat, even briefly, what is already known. Summaries strictly on the fulfilment of the proposed aims.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 6 Report

Comments and Suggestions for Authors

This paper is a very interesting case study on retrieval of three-dimensional green volume in urban green space from multi-source remote sensing data. Although some questions have not been well addressed, my suggestion is major revision at this stage.

 

1. Introduction: I suggest that authors should re-organize the the background and progress of the study, as the description of these two parts lacks sufficient logic.

 

2. Although the discussion section is very informative, I suggest that authors can give some potential policy suggestions for local goverment departments. 

 

3. The type of the reference should be revised according the rule of journal.

 

4. Some grammatical errors and irregular English writing can be found in the full text. I suggested that authors seek a native speaker to improve them. 

Comments on the Quality of English Language

Extensive editing of English language should be required. Please revise it. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

Extensive changes have been made and the paper is now acceptable for publication.

Comments on the Quality of English Language

Minor editing would be useful.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

Comments and Suggestions for Authors

It is meritorious that the authors have improved the ms. Both the Introduction and Material and Methods are very well written and meet the standards of the journal.

In the Results section, however, there are also texts, graphs and tables that 'tell the same story'.

Authors should be reminded that the Discussion section is not to mention previous studies but to compare their own results with others. Mention of previous studies will be made in the Introduction section.

Conclusions are fine.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 6 Report

Comments and Suggestions for Authors

Authors have revised the MS. I think it can be accepted at the current stage.

Author Response

Please see the attachment.

Author Response File: Author Response.docx