A Study on the Evolution of Forest Landscape Patterns in the Fuxin Region of China Combining SC-UNet and Spatial Pattern Perspectives
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis study introduces a novel method combining SC-UNet and Morphological Spatial Pattern Analysis (MSPA) for the extraction and analysis of forest landscapes in the Fuxin region of China. By integrating spatial and channel reconstruction convolution with the classical U-Net model, the research improves both the precision and speed of forest extraction from remote sensing images. The findings demonstrate significant advancements in the accuracy and efficiency of forest landscape monitoring, contributing valuable insights to the fields of remote sensing and ecological monitoring. The article is well-structured, logically coherent, and engages extensively with current literature, situating the research within the broader academic context. The references are appropriate and relevant, enhancing the study’s academic soundness and overall merit.
Author Response
Thank you for your comments, we appreciate your work!
Reviewer 2 Report
Comments and Suggestions for AuthorsThis work tries to improve the accuracy and speed of forest extraction from Sentinel-2 by proposing a new model SC-UNet, which combines two existing models, i.e., SCConv and U-Net, which have the advantages of low-redundancy and efficiency in remote sensing data processing and classification. The model was then validated with several areas selected within the northeastern provinces of China. Based on the results of the proposed model, the trends and evolutionary patterns of the forest in the Fuxin Region of China between the two time stamps 2019 and 2023 are then analyzed. The general logistics are reasonable. However, the significance of this work is not clear, and there are some key issues that need to be clarified and improved.
General comments:
There are many times mentioning "large-scale" in this work. Considering the study area is actually not that "large" in the remote sensing field, I suggest the authors avoid this misleading word.
L21-22, the sentence is misleading. The reader may misunderstand that the work has processed the whole northeastern China.
The authors mentioned the time period 2019-2023(L25) and 2016-2023 (L80). Which one, exactly? The authors seem to have only checked the two years 2019 and 2023. I suggest not to mention "high temporal resolution" (L41) at all. Moreover, we shouldn't say "trend" with only two time stamp data. "change" may be a better word.
L353-355, by saying "we also constructed a large-scale, high intraclass diversity forest land dataset using Sentinel-2 10-m resolution multispectral satellite remote sensing imagery, which covers most of northeastern China", do the authors mean this is a quite large binary image database that can be used for training the models? Does it mean all the 9220 samples (L262)? How was this training dataset created, manually?
The background on the importance of the topic "Evolution of Forest Landscape Patterns" is not clearly provided in the Introduction part nor key related literatures cited.
The same problem exists for the statement "ecosystem restoration" (L29-30). How can this work contribute to assessment of ecosystem restoration?
It lacks a figure in Section 2.1 to show the geolocation and basic geoinformation of the area. L101, "The Fuxin area (41°17′–45°24′N, 116°21′–120°58′E)" is not consistent with Fig 4-7.
Are the binary images in Table 2 ground truths, representing forests (white) and non-forests (black)? If yes, I suggest changing the position of the groundtruth and the images and stating clearly in the table caption what those images are. Because at first sight, readers would misunderstand that you get groundtruth from those images. For example, it is not possible to get the binary images from the cloudy images in the fourth column.
I doubt the results of U-Net in Table 5. It is weird that there are not any speckles in the red boxes, that seems to be a misusage of the model. Large blocks/spots are also disappearing, that is unusual in regular image processing.
L344, "Changwu sandy area" should be clearly marked on the map.
Minors:
L11-13, there are three "more" in the first sentence that are not clear and easily understood. Please rewrite it.
L18, full name of "IoU" should be used here
L123, [41-45] These citations seem irrelevant to the sentence
Figure 1, GRU->CRU
Eq.2-3 are not right. The whole right parts are in the subscript
Comments on the Quality of English LanguageThere are English grammar problems that need to be checked.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear Authors,
The study entitled by "A Study on the Evolution of Forest Landscape Patterns in the Fuxin Region of China Combining SC-UNet and Spatial Pattern Perspectives" has been reviewed, and in general this is a well-written study, and my specific comments are as follows.
1 The detailed information should be added, such as the used bands, the number of Sentinel-2 imagery.
2 Regarding the maps in this study, I think the additional map showing the changes over 2019-2023 is highly needed.
3 Figures 4,5,6,7 are coarse, the resolution needs improvement. Moreover, these maps need the compass and scale bar.
4 The study area should be mapped using a figure for better demonstration.
5 Regarding the spatial pattern of the forest change over 2019-2023, the potential factors, such as natural factors and human activities, should be discussed in the manuscript, please refer the following study for discussion of various factors of forest changes.
Li, L., Zhu, L., Xu, N., Liang, Y., Zhang, Z., Liu, J. and Li, X., 2022. Climate Change and Diurnal Warming: Impacts on the Growth of Different Vegetation Types in the North–South Transition Zone of China. Land, 12(1), p.13.
6 Is there any statistical data available for comparison？
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsTwo minor comments:
1. In figure 1, there is a misleading format between the two region names. At first sight, the two regions are on the same level, with one representing the left half and the other the right.
2. About Eqs. 2–3, the authors have misunderstood my comment. I meant that the whole equation is a subscript except for the letter "Y". You need to move the right half, starting from "=", out of the subscript to a normal format.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear Authors,
The manuscript has been largely improved, only some little issues. First, some references are with different styles, some changes are needed. Second, because sentinel-2 are available on the google earth engine platform, the use of GEE is highly recommended in the future.
references
Advancing the mapping of mangrove forests at national-scale using Sentinel-1
Google Earth Engine and artificial intelligence (AI): a comprehensive review
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
