A New Technique for Impervious Surface Mapping and Its Spatio-Temporal Changes from Landsat and Sentinel-2 Images

Round 1

Reviewer 1 Report

The current manuscript is facing a number of problems and need to be a major revision. For example, I list some of them here.

Line 13: The abstract is not written well.

Line 82: Something wrong here. The main function of Atmospheric correction is not like that. Please correct it.

Line 87: I cannot agree with this conclusion. They are demonstrated one of the most accuracy methods in atmospheric correction since they require a huge input data. DOS, frankly speaking is not the most accuracy technique for the correction.

Line 163: Why only 5 scenes are selected?

Line 238: I recommend the authors consider using ACOLITE for the atmospheric correction. Such method could apply for both Landsat and Sentinel.

Line 357: I am doubting this result. Very high surface reflectance observed due to the atmospheric correction failed ?

See an example:

https://www.researchgate.net/publication/352403454_An_Effective_Water_Body_Extraction_Method_with_New_Water_Index_for_Sentinel-2_Imagery

Line 358: As I said, using two different method for atmospheric correction of the same purpose is not appropriate. You should apply the same approach for atmospheric correction. Hence, you and further analysis the data.

Line 373: How about the AA for the atmospheric correction? You mentioned that this step is important in the image analysis, but there is no final assessment for the use of DOS and Sen2Cor.

Line 458: Table-based illustration will be much better than the writing like this.

Line 579: What is the white color in this map?

Line 654: It is too difficult to catch up all the information here. Please re-write this paragraph.

 

Details shown in the attached file.

Comments for author File: Comments.pdf

Author Response

We are very encouraged by the insightful comments from the editor and the anonymous referees. They provided many helpful suggestions to clarify our ideas, to correct language errors, and to improve the figures. We have substantially revised the manuscript based on these comments. The main revisions are listed as follows:

 

Reviewer 1

 

Reviewer’s Comment: Line 13: The abstract is not written well.

Responses from authors: Thank you for the constructive comments. The abstract has been modified.

Reviewer’s Comment: Line 82: Something wrong here. The main function of Atmospheric correction is not like that. Please correct it.

Responses from authors: Thank you for the constructive comments. We corrected it. The purpose of atmospheric correction is to obtain precise surface reflectance values by eliminating the influences of the atmosphere, such as scattering and absorption, on remotely sensed imagery. By applying atmospheric correction techniques, these effects can be accounted for, resulting in corrected images that more accurately depict the characteristics and circumstances of the earth's surface.

Reviewer’s Comment: Line 87: I cannot agree with this conclusion. They are demonstrated one of the most accuracy methods in atmospheric correction since they require a huge input data. DOS, frankly speaking is not the most accuracy technique for the correction.

Responses from authors: Thank you for the constructive comments. We corrected it. The physically based models, such as 6S, LOWTRAN, and MODTRAN, produce very high surface reflectance accuracy. However, the models are complicated and require a lot of input parameters from the in situ field atmospheric data received at the time of remote sensing data collecting.

Reviewer’s Comment: Line 163: Why only 5 scenes are selected?

Responses from authors: Thank you for the constructive comments. The two sentinel-2 images (89-T50RPN and 89-T50RNN) in Table 1 need be mosaicked to create a complete image encompassing Xiamen City.

Reviewer’s Comment: Line 238: I recommend the authors consider using ACOLITE for the atmospheric correction. Such method could apply for both Landsat and Sentinel.

Line 358: As I said, using two different method for atmospheric correction of the same purpose is not appropriate. You should apply the same approach for atmospheric correction. Hence, you and further analysis the data.

Responses from authors: Thank you for the constructive comments. We appreciate your suggestion on using the ACOLITE method for atmospheric correction of Landsat and Sentinel-2 images. We will carefully considered the ACOLITE method in future work. However, we have decided to continue using the DOS method for Landsat images and the sen2cor model for Sentinel-2 images in current work.

There are several reasons for this decision. Firstly, we have already validated our results using the DOS and sen2cor methods, and changing methods at this point would require repeating much of our analysis. Secondly, we have found that the accuracy of our results using the current methods are acceptable for our research objectives. Finally, we believe that using different atmospheric correction methods for different datasets may introduce additional sources of error and variability that could complicate our analysis.

We understand that the ACOLITE method may have some advantages over the current methods we are using, but we believe that our current approach is appropriate for our study. We hope that you will agree with our decision and find our explanation satisfactory.

Reviewer’s Comment: Line 357: I am doubting this result. Very high surface reflectance observed due to the atmospheric correction failed? See an example: https://www.researchgate.net/publication/352403454_An_Effective_Water_Body_Extraction_Method_with_New_Water_Index_for_Sentinel-2_Imagery

Responses from authors: Thank you for the constructive comments. We made appropriate modifications in our paper. Firstly, we have modified the word "Vegetation" to "Forestland" in our figure 5.

In addition, we compared our study with the reference study conducted by Jia et al. Our analysis shows that the reflectance values for sea in our study are similar to those (general water) observed in Jia's study. Specifically, the reflectance of water gradually decreases from TM1 to TM7 for Landsat 8 image, with a maximum value less than 0.2. The mean NDWI value for general water is approximately 0.55 in our study, which is comparable to the mean NDWI value (0.4) observed in Jia's study.

However, there are some differences between our study and Jia's study. For instance, Jia used TOA reflectance values, while we used surface reflectance values. Additionally, Jia calculated the mean NDWI value for general water in the whole study area, whereas we used the mean NDWI value for some samples. Nevertheless, the results obtained in both studies are similar and not significantly different.

Reference: Jiang, W., Ni, Y., Pang, Z., Li, X., Ju, H., He, G., ... & Qin, X. An effective water body extraction method with new water index for sentinel-2 imagery. Water, 2021, 13(12), 1647

Reviewer’s Comment: Line 373: How about the AA for the atmospheric correction? You mentioned that this step is important in the image analysis, but there is no final assessment for the use of DOS and Sen2Cor.

Responses from authors: We appreciate the feedback and would like to address the concern regarding the final assessment of DOS and Sen2Cor in our study. Please See Section 3.3.4 Assessment of DOS and Sen2Cor. Compared to raw data, the atmospherically corrected images are visually clearer because the atmosphere's effects have been well removed in the new images. To quantify the comparison between raw images and corrected images by atmospheric correction, we used the coefficient of variation (CV) as a measure, which was proposed by Yang and Lo (2000). A higher CV value indicates a more dispersed distribution, indicating that the images with larger CV values can lead to better remote sensing classification.

In Figure S of the supplemental file, we included the CV values computed for each band (excluding thermal band) for different images using atmospheric correction methods. The figure also shows the average of these measures for all bands for each year from 2009 to 2021. For comparison purposes, we computed the same measures for the raw images. As shown in Figure S, all CV values of corrected images are greater than those of raw images. This suggests that our atmospheric correction methods were very effective in providing important data for subsequent ISA mapping and landcover changes analysis. We hope this additional information addresses any concerns regarding the effectiveness of our atmospheric correction methods.

Reference: Yang, X., Lo, C. P. Relative radiometric normalization performance for change detection from multi-date satellite images. Photogramm. Eng. Rem. Sens. 2000, 66(8): 967-980.

Reviewer’s Comment: Line 458: Table-based illustration will be much better than the writing like this.

Responses from authors: Thank you for the constructive comments. We improved the expression about the paragraph. To avoid duplication, a table-based illustration has not been included in addition to Figure 8, which already shows the comparison of ISA classification accuracy by OA, KC and MICE.

Reviewer’s Comment: Line 579: What is the white color in this map?

Responses from authors: Thank you for the constructive comments. I modified the legend in the map. The white color in this map represents Non-ISA areas.

Reviewer’s Comment:  Line 654: It is too difficult to catch up all the information here. Please re-write this paragraph.

Responses from authors: Thank you for the constructive comments. We rewrite this paragraph. Please see the modified paper.

Reviewer 2 Report

In this paper, a classification method based on decision tree approach and some spectral indexes is developed. It seems that the proposed framework is new in this study. The paper is interesting but is not mature enough for publication. Please consider following comments:

·         Title can be presented in a better form. For example a coastal city is not necessary to mention in this form.

·         Abstract is supported by many results. I think it can be reduced. Please mention important results.

·         Figure 1: please provide title with more descriptions.

·         Line 176 is not meaningful.

·         Line 180: please mention reasons behind selecting steps of the method. I did not understand why these atmospheric correction methods are employed.

·         Figure 2: what is difference between image processing and atmospheric corrections? Moreover, Images based on Sen2Cor model is not connected to any box, why?

·         Section 3.2.1: full of equations! Do you think that is it necessary for readers?

·         Table2: again I think it is not necessary for this study.

·         Figure3: please use colors similar to legend of maps.

·         Section 3.3.4 TB of BT?

·         Please mention limitations of this study. Moreover, source of errors should be mentioned and discussed in the text.

 

Author Response

Re: A New Technique for Impervious Surface Mapping and its Spatio-temporal Changes for a Coastal City, China Using Landsat and Sentinel-2 Images

 

We are very encouraged by the insightful comments from the editor and the anonymous referees. They provided many helpful suggestions to clarify our ideas, to correct language errors, and to improve the figures. We have substantially revised the manuscript based on these comments. The main revisions are listed as follows:

 

Reviewer 2

In this paper, a classification method based on decision tree approach and some spectral indexes is developed. It seems that the proposed framework is new in this study. The paper is interesting but is not mature enough for publication. Please consider following comments:

• Reviewer’s Comment:Title can be presented in a better form. For example, a coastal city is not necessary to mention in this form.

Responses from authors: Thank you for the constructive comments. The title has been modified as A New Technique for Urban Impervious Surface Mapping and its Spatio-temporal Changes from Landsat and Sentinel-2 Images

Reviewer’s Comment: Abstract is supported by many results. I think it can be reduced. Please mention important results.

Responses from authors: Thank you for the constructive comments. The abstract has been modified. Please the new abstract.

Reviewer’s Comment: Figure 1: please provide title with more descriptions.

Responses from authors: Thank you for the constructive comments. The new title of Figure 1 is “The location of Xiamen City, the study area consisting of four new urban clusters in bays and a central city on Xiamen Island, which is illustrated by the undulating topography.”

 

Reviewer’s Comment:   Line 176 is not meaningful.

Responses from authors: Thank you for the constructive comments. The line has been removed.

 

Reviewer’s Comment:  Line 180: please mention reasons behind selecting steps of the method. I did not understand why these atmospheric correction methods are employed.

Responses from authors: Thank you for the constructive comments. The purpose of atmospheric correction is to obtain precise surface reflectance values by eliminating the influences of the atmosphere, such as scattering and absorption, on remotely sensed imagery. By applying atmospheric correction techniques, these effects can be accounted for, resulting in corrected images that more faithfully depict the characteristics and circumstances of the earth's surface. The study applied Level-1C products of Landsat and Sentinel-2 images. It is necessary to correct the atmospheric effects.

 

Reviewer’s Comment:  Figure 2: what is difference between image processing and atmospheric corrections? Moreover, Images based on Sen2Cor model is not connected to any box, why?

Responses from authors: Thank you for the constructive comments. Figure 2 has been redrawn. The term "image processing" referenced in Figure 2 has been updated to "image preprocessing". Some common tasks involved in remote sensing image preprocessing include image clipping, radiometric calibration, atmospheric correction, geometric correction and so on. Atmospheric correction is an important part of image preprocessing process preprocessing. In addition to atmospheric correction,

 

Reviewer’s Comment: Section 3.2.1: full of equations! Do you think that is it necessary for readers? Table2: again I think it is not necessary for this study.

Responses from authors: Thank you for the constructive comments. We have simplified three image-based atmospheric correction methods for both Landsat 5 and Landsat 8 images. Our simplified methods enable faster computation compared to the original DOS model. It is necessary to maintain important equations to ensure that readers understand the derived processes. Despite this, we have reduced some equations in the paper for clarity.

 

Reviewer’s Comment:  Figure3: please use colors similar to legend of maps. ·

Section 3.3.4 TB of BT?

Responses from authors: Thank you for the constructive comments. Figure 3 has been set colors similar to legend of map in Figure 6.

TB represents Bright temperature. The article has replaced all instances of the acronym BT to TB.

Reviewer’s Comment: Please mention limitations of this study. Moreover, source of errors should be mentioned and discussed in the text.

Responses from authors: Thank you for your valuable comments. We have added the limitations and source of errors. Nevertheless, this proposed method has some limitations. First, the parameters of the model cannot be calculated automatically. Thus, it is crucial to combine FDTC with the effective remote sensing image segmentation algorithm to obtain the parameters in a timely manner. Second, because FDTC uses thermal infrared band to distinguish wetland from high-rise buildings, some mixed pixels are generated when FDTC is applied to high-resolution Sentinel-2 images. To overcome this problem, the thermal infrared images must be refined by downscaling with land surface temperature. In addition, we utilized the DOS and sen2cor methods to eliminate the effects of the atmosphere on Landsat and Sentinel-2 images. However, to compare changes in FDTC parameters resulting from various atmospheric correction methods, future research should consider investigating other techniques such as ACOLITE. Finally, this study uses FDTC only in one coastal city, so further experimentation should consider other areas with different environmental conditions using more medium- and high-resolution images.

Round 2

Reviewer 1 Report

Thank you for the hard work. I think the revised version is good enough to be published in Sustainability.

Reviewer 2 Report

-