Land Use/Cover Change Prediction Based on a New Hybrid Logistic-Multicriteria Evaluation-Cellular Automata-Markov Model Taking Hefei, China as an Example
Round 1
Reviewer 1 Report (New Reviewer)
Good and valuable research has been done; however, the manuscript is not well organized. It could be accepted after the major corrections and re-evaluation.
The abstract needs to be rewritten. It is more appropriate to provide a bit with more details and within the limit of allowed words.
Authors are recommended to emphasis the novelty and significance of the study in more detail.
The introduction section is poor and haphazard. The author enlisted a few scholars' work without a linkage with the key issues of the study. The literature cited in this section is insufficient. The author should conduct more.
Considering the importance of part “Materials and Methods”, it is recommended to provide more complete explanations in this regard.
Authors are recommended to discuss the obtained results with literature in more detail.
The results and discussion sections are also poor and haphazard, there is no planning for presenting the findings properly. The description lacks coherence due to procedural weakness. The authors are advised to revise the whole sections by following a new methodology.
The conclusion section is not structured. It should be revised by adding key findings, recommendations, and practical implications.
Minor editing of English language required.
Author Response
Response to the reviewer’s comments:
To Reviewer #1
- Good and valuable research has been done; however, the manuscript is not well organized. It could be accepted after the major corrections and re-evaluation.
Response: Thank you for your positive comments. We have carefully reorganized the paper according to your suggestion in order to improve further the paper.
- The abstract needs to be rewritten. It is more appropriate to provide a bit with more details and within the limit of allowed words.
Response: Thank you for the suggestion. The abstract of this paper has been rewritten, which is more in line with the topic of the paper.
- Authors are recommended to emphasis the novelty and significance of the study in more detail.
Response: Thank you for your comments. It is highlighted in Conclusion in revision.
- The introduction section is poor and haphazard. The author enlisted a few scholars' work without a linkage with the key issues of the study. The literature cited in this section is insufficient. The author should conduct more.
Response: Thank you for your comments. We have rewritten the Introduction section to make it more coherent. The references that are not closely related to this study are screened out, and those closely related to the key issues of this study are added to ensure that the references are closely combined with the research content.
- Considering the importance of part “Materials and Methods”, it is recommended to provide more complete explanations in this regard.
Response: Thanks for your comments. We have added a description of the principles of each model in the "Materials and Methods" section, making the paper more rigorous and complete. Please check our revision.
- Authors are recommended to discuss the obtained results with literature in more detail.
Response: Thank you for your suggestions. As suggested, we have added relevant literature to this study and compiled a more detailed discussion of the results.
- The results and discussion sections are also poor and haphazard, there is no planning for presenting the findings properly. The description lacks coherence due to procedural weakness. The authors are advised to revise the whole sections by following a new methodology.
Response: Thank you for your suggestion. The methodology, results and discussion are all restructured. Please check our revision.
- The conclusion section is not structured. It should be revised by adding key findings, recommendations, and practical implications.
Response: Thank you for your suggestion. We reorganize the structure of the conclusion and rewrote it according to the key findings, recommendations, and practical implications.
Author Response File: 
Author Response.docx
Reviewer 2 Report (New Reviewer)
The research is very interesting for all hydrological community in the world.
I have a minor observation:
All the document is presented in EDITION/REVISION MODE. A lot of words and paragraphs (even graphics) are crossed-out and have corrected version besides them. It was difficult for me to read the whole document. Please present a clean version of your paper to read it better.
- What is the main question addressed by the research?
- Do you consider the topic original or relevant in the field? Does it address a specific gap in the field?
- What does it add to the subject area compared with other published material?
- What specific improvements should the authors consider regarding the methodology? What further controls should be considered?
- Are the conclusions consistent with the evidence and arguments presented and do they address the main question posed?
- Are the references appropriate?
- Please include any additional comments on the tables and figures.
Author Response
Response to the reviewer’s comments:
To Reviewer #2
- The research is very interesting for all hydrological community in the world.
Response: Thank you for your positive comments. We will continue to improve the paper based on your suggestions, hoping for better quality.
- I have a minor observation:
All the document is presented in EDITION/REVISION MODE. A lot of words and paragraphs (even graphics) are crossed-out and have corrected version besides them. It was difficult for me to read the whole document. Please present a clean version of your paper to read it better.
Response: Thank you for the suggestion. We will accept your suggestion and submit both change-tracking and clean versions.
- What is the main question addressed by the research?
The main question addressed by the research is the possibility of increase the prediction of Land Use / Land Cover change in a 5-year period of time according to the paper (2020 - 2025). The model suggested by the authors predicts better than other models.
Response: Thank you for your valuable comments.
- Do you consider the topic original or relevant in the field? Does it address a specific gap in the field?
Nowadays, the topic is relevant in the field. In developing countries, many authorities want decision-making tools regarding flooding, and LULC change predictions could help in the decision-making process. Attempts in this kind of change prediction are not new research, but the paper approach is relatively new.
Response: Thank you for your comments.
- What does it add to the subject area compared with other published material?
The use of Logistic-MCE-CA-Markov (LMCM) model is the addition to the subject area compared with other published material.
- What specific improvements should the authors consider regarding the methodology? What further controls should be considered?
Regarding methodology, there is no comment about it. However, there is a word that should be corrected (lines 16 and 68). It is written "Celluar". I think it should be "Cellular" There is also a highlighted line (line 155) that should not be highlighted, correct?
Response: Thank you for your comments and suggestion. No, it shouldn’t.
- Are the conclusions consistent with the evidence and arguments presented and do they address the main question posed?
Yes, they are.
Are the references appropriate?
Yes, they are.
Please include any additional comments on the tables and figures.
I do not have any additional comment on tables and figures.
Response: Thank you for your time and suggestions.
Author Response File: Author Response.docx
Reviewer 3 Report (New Reviewer)
The introduction is OK, but a little more clarity is required the explain why a new model is required, and why current methods are not adequate. Are there new challenges current methods are insufficient for, are there noted gaps in the literature that are not currently able to be tackled using the current methods. The paper does a little but of this currently, but some more applied examples would be helpful.
Why was this case study regions chosen for the study - are there specific characteristics that were important here?
Although the methods make sense to me, I think a little more explanation for what each step is trying to do with reference to the overall research design would be helpful - just a sentence or two.
Figure 3 - suggest using different colours than red/blue/green to provide more accessibility for colour blind individuals.
Figure 5 needs additional narrative explanation to aid interpretation
Because I am not convinced yet, based on the introduction, that there is a problem to solve, I am also not convinced this new model methods is actually contributing all that much - this is not to say it doesn't, I just don't think the authors have made that case satisfactorily.
Not too bad - would benefit from an english language edit.
Author Response
Response to the reviewer’s comments:
To Reviewer #3
- The introduction is OK, but a little more clarity is required the explain why a new model is required, and why current methods are not adequate. Are there new challenges current methods are insufficient for, are there noted gaps in the literature that are not currently able to be tackled using the current methods. The paper does a little but of this currently, but some more applied examples would be helpful.
Response: Thank you for your suggestion. We have rewritten the introduction section. Firstly, we screened the references and added those that are more consistent with the key issues of this study, as well as application examples. In addition, we supplemented the challenges of the current methods and the necessity of constructing a new model.
- Why was this case study regions chosen for the study - are there specific characteristics that were important here?
Response: Thank you for raising this issue. There are three main reasons for choosing Hefei as the research area: First, Hefei is the capital city of Anhui Province, with an area of more than 10000 km² and composed of various land use types. Second, Hefei's GDP, population, urbanization rate and other aspects have increased significantly from 2010 to 2020, which are likely to be bound to affect the process of land use change, and this may pose a challenge for the application of the available models. Finally, the research data of Hefei is complete and available, so we chose this area. Thank you for raising this question.
- Although the methods make sense to me, I think a little more explanation for what each step is trying to do with reference to the overall research design would be helpful - just a sentence or two.
Response: Thank you for your comments. Following your suggestion, we have provided more detailed supplements to the steps for the overall research design in the beginning of Section 2 Materials and Methods.
- Figure 3 - suggest using different colours than red/blue/green to provide more accessibility for colour blind individuals.
Response: Thank you for your comments. It is impressive for your attention to the reading experience of colour blind individuals, but due to the fact that the colors of all images in this paper are uniformly matched based on reference to existing literature, and the time for this revision is quite urgent, if uniformly modified, it may lead to inaccurate color matching. So, we have provided a method for identifying colors designed for colour blind individuals in Appendix 4, hoping that they can better read this paper.
- Figure 5 needs additional narrative explanation to aid interpretation.
Response: Thank you for your comments. According to your suggestion, we have written the narrative explanation of Figure 5 in more detail.
- Because I am not convinced yet, based on the introduction, that there is a problem to solve, I am also not convinced this new model methods is actually contributing all that much - this is not to say it doesn't, I just don't think the authors have made that case satisfactorily.
Response: Thank you for your comments. We have rewritten the paper to better express our research work. Indeed, the new model we are currently proposing still has some shortcomings, such as the difficulty in predicting land use type changes caused by policy impacts. We will continue to delve into such issues in future research with the hope of making progress.
Author Response File: Author Response.docx
Reviewer 4 Report (New Reviewer)
Editor-in-Chief
Land
# land-2618478
Manuscript Title: LUCC prediction based on a new hybrid LMCM model taking Hefei, China as an example
Having carefully evaluated the manuscript, I would like to commend the authors for their thoughtful and rigorous work. I would like to offer some feedback for the authors' consideration during the revision process, which may improve the strength of the paper. Moreover, I believe that the manuscript is well organized. Therefore, a MINOR revision is recommended for the manuscript. Details of the comments are as follows:
Introduction:
- Please provide further elaboration on the various approaches to LUCC prediction.
- Prior to stating the objective, it is essential to highlight the significance of conducting research in the study area and its characteristics for selection as the study area.
Research Methodology:
- Provide a clearer explanation regarding the selection of the vegetation indices.
- On what criteria was the selection of the year 2025 as the target year based?
- Has the IDW (Inverse Distance Weighting) method been suitable for spatial interpolation of the influential variables? Please provide additional explanations on this matter.
Results:
- In the interpretations, provide some explanations regarding the inter-dependency among variables affecting LUCC change.
Conclusions:
- The limitations of the research and source of uncertainties would be better to mention in the conclusion section.
Author Response
Response to the reviewer’s comments:
To Reviewer #4
- Having carefully evaluated the manuscript, I would like to commend the authors for their thoughtful and rigorous work. I would like to offer some feedback for the authors' consideration during the revision process, which may improve the strength of the paper. Moreover, I believe that the manuscript is well organized. Therefore, a MINOR revision is recommended for the manuscript.
Response: Thank you for your comments and decision. We have carefully revised the paper according to your suggestions, hoping to further improve the quality of the paper.
- Details of the comments are as follows:
Introduction:
- Please provide further elaboration on the various approaches to LUCC prediction.
Response: Thank you for your comments. According to your suggestion, we rewrote the introduction, supplemented the references closely related to this study, and further explained various approaches to LUCC prediction.
- Prior to stating the objective, it is essential to highlight the significance of conducting research in the study area and its characteristics for selection as the study area.
Response: Thank you for your comments. But the problematic issues were to mentioned in Subsection 2.1 (Study area). Logically, it seems better that we emphasized this in Study area following an introduction of the study area and then raised the problems there, and the need for a simulation study.
- Research Methodology:
- Provide a clearer explanation regarding the selection of the vegetation indices.
Response: Thank you for your comments. Following your suggestion, we have added the reasons for choosing these three vegetation indices in '2.2.2. Processing Procedures'. It goes like this “It is known that EVI can effectively reduce soil and atmospheric impacts [69], and GDVI is of higher sensitivity and wider dynamic range to low vegetated land cover such as bareland, urban areas, grasslands or rangelands than other vegetation indices [67], and NDVI lies in between EVI and GDVI, a compromise of the two indices. Thus, utilization of the three indices may contribute to identify land cover with more detail.”
- On what criteria was the selection of the year 2025 as the target year based?
Response: Thank you for your comments. We have classified the land use types of remote sensing images in 2010, 2015 and 2020 with a time interval of 5 years, so we first thought of the prediction of 2025. Of course, our research is not limited to the prediction of 2025. With sufficient data, we can predict land use change for a longer time.
- Has the IDW (Inverse Distance Weighting) method been suitable for spatial interpolation of the influential variables? Please provide additional explanations on this matter.
Response: Thank you for your comments. Actually, both kriging and IDW (inverse distance weighting) are frequently applied interpolation approaches. Though Milillo and Gardella (2008) found that ordinary kriging is more accurate than IDW in retention of original image features, Spokas et al. (2003) and Gong et al. (2014) have shown that IDW is superior to kriging in estimation of landfill methane flux and groundwater arsenic concentrations whereas Munyati and Sinthumule (2021) found that for forest, kriging has higher correlation of tree density with NDVI than IDW while for woodlands, IDW has higher correlation than kriging in estimating tree density. Thus, whether kriging or IDW interpolation performs better is case dependent. For simplicity, we selected IDW for interpolation of these factors. We have provided additional explanations on this in '2.2.2. Processing Procedures '.
References
Gong, G., Mattevada, S., O’Bryant, S.E. Comparison of the accuracy of kriging and IDW interpolations in estimating groundwater arsenic concentrations in Texas. Environmental Research, 2014. 130: 59-69. https://doi.org/10.1016/j.envres.2013.12.005.
Milillo, T.M., Gardella, J. Spatial Analysis of Time of Flight−Secondary Ion Mass Spectrometric Images by Ordinary Kriging and Inverse Distance Weighted Interpolation Techniques. Analytical Chemistry, 2008, 80(13):4896-905. DOI: 10.1021/ac702640v.
Munyati,C., Sinthumule N.I. Comparative suitability of ordinary kriging and Inverse Distance Weighted interpolation for indicating intactness gradients on threatened savannah woodland and forest stands. Environmental and Sustainability Indicators,2021,12,100151. https://doi.org/10.1016/j.indic.2021.100151.
Spokas, K., Graff, C., Morcet, M., Aran, C. Implications of the spatial variability of landfill emission rates on geospatial analyses. Waste Management, 2003, 23(7): 599-607. https://doi.org/10.1016/S0956-053X(03)00102-8.
- Results:
- In the interpretations, provide some explanations regarding the inter-dependency among variables affecting LUCC change.
Response: Thank you for raising this issue. Our test revealed that VIF of all the remained variables is < 10, and this indicates that there is no collinearity or inter-dependency among the factors. The previous Tables 5 and 6 (indicating VIF) were moved to Subsection 2.3 Methods as new Tables 3 and 4 in revision. Two factors such as temperature and rainfall were removed after collinearity diagnosis.
- Conclusions:
- The limitations of the research and source of uncertainties would be better to mention in the conclusion section.
Response: Thank you for your comments. Following your suggestion, we have added the limitations of the research and source of uncertainties in the conclusion section.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report (New Reviewer)
Thank you very much for your revised manuscript. No more comments on this paper.
Minor editing of English language required.
Reviewer 2 Report (New Reviewer)
Good final work.
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 aims at simulating LULC by proposing a new hybrid LMCM model for better estimation of landcover features in Hefei, Anhui Province, China. The paper seems interesting and in the aims of the Journal. However, following are a few comments for further improvement.
1. The writing needs improvement. Also, having a native English speaker to go over the manuscript should be warranted.
2. References in the article should be rechecked for mistakes and recommended citation as well as referencing style should be adopted.
3. Please add up-to date and relevant references to support the explanation/details provided for study area in section 2.1
4. The introduction and other sections should be re-reviewed to avoid typing errors such as, date formats. Also, full form of the abbreviation should be written well before the short-termed abbreviations. Like the term GDP should be clearly elaborated and abbreviated scientifically.
5. After elaborating similar studies covering the targeted region novelty of the study should be clearly emphasized.
6. Table format should be consistent and adopted as per the journal requirements. Also, try adjusting the Table 8 on next page.
7. See lines 258, 259, 260, 451, and 477 to address associated error in referencing.
8. Based on your findings/conclusions, suggest a few areas for future research.
9. Avoid references older than 5 years (>2018)
10. There are many methods and tools available to perform the spatial interpolation. Justify the reason of selecting/ using IDW approach. Why did not the authors utilize kriging or trend surface estimation? Please explain.
The writing needs improvement. Also, having a native English speaker to go over the manuscript should be warranted.
Author Response
Response to the reviewer’s comments:
To Reviewer #1
- The writing needs improvement. Also, having a native English speaker to go over the manuscript should be warranted.
Response: Thank you for your comments. We have carefully checked and improved the English writing in the revised manuscript. In addition, the expression of the manuscript has been improved with the help of a native English speaker.
- References in the article should be rechecked for mistakes and recommended citation as well as referencing style should be adopted.
Response: Thank you for the suggestion. References and their citation have been revised in terms of the journal's requirement.
- Please add up-to date and relevant references to support the explanation/details provided for study area in section 2.1.
Response: Thanks for the references, which are now included in the revised manuscript, which are listed as references 34 and 36.
- The introduction and other sections should be re-reviewed to avoid typing errors such as, date formats. Also, full form of the abbreviation should be written well before the short-termed abbreviations. Like the term GDP should be clearly elaborated and abbreviated scientifically.
Response: Thank you for raising this issue. Grammar issue and type errors together with the abbreviation involved in this article were revised carefully according to your suggestions, and added full form of the abbreviation, such as GDP (Gross Domestic Product) and COST(Cosine Theta). Please note that the cosine function of the zenith angle (θ) was regarded as an approximation of the upwelling transmittance from ground to sensor by Chavez in 1996 and this is a key for COST model to conduct image-based atmospheric correction.
- After elaborating similar studies covering the targeted region novelty of the study should be clearly emphasized.
Response: Thanks for your comments. The novelty of the new hybrid model LMCM different from the existing models was highlighted in Discussion (4.2).
- Table format should be consistent and adopted as per the journal requirements. Also, try adjusting the Table 8 on next page.
Response: Thank you for your suggestions. We have revised tables, in particular, Table 8 and replaced "proportion" with "change between 2020 and 2025" and "average annual change" to further illustrate future land-use change trend.
- See lines 258, 259, 260, 451, and 477 to address associated error in referencing.
Response: Thank you for raising this issue. We re-examined the references and formatting them into the required styles of journal in revision.
- Based on your findings/conclusions, suggest a few areas for future research.
Response: Thank you for your suggestion. Since the LMCM model performs better than other hybrid models, we would suggest that a further study would be focused on regional and even national studies by upscaling this model for simulating LUCC of future. Hence, our research may provide an added value in guiding urban planning and spatial planning of territory. This is added in Conclusions.
- Avoid references older than 5 years (>2018).
Response: Thank you for raising this but without the works of pioneers, there should not be the achievements of successors. To remove all those older than 5 years is not a good way of scientific communication. We prefer to keep them, at least, those are essential for introduction to our research.
- There are many methods and tools available to perform the spatial interpolation. Justify the reason of selecting/ using IDW approach. Why did not the authors utilize kriging or trend surface estimation? Please explain.
Response: Thank you for your question. The main reason for choosing IDW in this study lies in the fact that the basic idea of IDW is that the closer the unknown point to the data point, the closer the attribute value to the attribute value of the data point. Therefore, the weighted average of the attribute value of the unknown point by distance and weight can obtain an accurate estimate, which can meet our study. Regarding kriging or trend surface estimation, each has its advantage but this is not the purpose to compare their advantages of this paper. Probably, we may specifically focus an analysis on this comparison in future.
Author Response File: Author Response.docx
Reviewer 2 Report
By reviewing the existing widely adopted land use/cover change (LUCC) models and identifying their limitations, this study put forward an integrated model (LMCM) to address these limitations, which provides a more accurate approach for LUCC study and urban planning practices. The city of Hefei in China was taken as case study to test the LMCM model and its land use/cover change in 2025 was also predicted and analyzed. I think this is a good manuscript. The topic to be presented in the manuscript is relevant and well fits the scope of the journal. In general, the objective and presentation form of the manuscript are clear, and the writing language is fine, making the results convincing. The presentation structure is logical and follows a step-by-step sequence. However, a few aspects can still be adjusted and improved.
1) There are a lot of abbreviations in the manuscript, it would be better to provide a list of those abbreviations at the beginning of the article or show them in the Appendix.
2) Type of data used. Information in Table 1 should be consistent with text description. Remote sensing and terrain are categorized as two data types, while they were combined as one in the text descriptions. Also, Field investigation data was not presented in Table 1.
3) Regarding the sub-section “3.4. Land use pattern of 2025”, the result of this predication analysis should be further elaborated by highlighting the trend of changes as well as rates. It is also suggested elaborating further Table 8 by adding comparison between the situation of years 2020 and 2025.
4) As the study also aims to provide scientific support for spatial planning of territory in Hefei City, the specific policy implications for this city should also be discussed.
5) As the integrated model LMCM model excludes experts’ knowledge, authors should clarify why this dimension is not important and what are the limitations when involving experts’ knowledge in terms of land use/cover change analysis and prediction.
6) The in-text citation style needs to be carefully checked and adjusted, as the cited numbers are mixed with the texts, while some of the citations are missing.
No comments for the English language.
Author Response
Response to the reviewer’s comments:
To Reviewer 2
- There are a lot of abbreviations in the manuscript, it would be better to provide a list of those abbreviations at the beginning of the article or show them in the Appendix.
Response: Thank you for your suggestions. We have re-checked the abbreviations in the manuscript and the abbreviations which were missing in the initial submission such as GDP and COST were fully spelled out. Apart from the text, all abbreviations have been clearly presented in Appendix 1.
- Type of data used. Information in Table 1 should be consistent with text description. Remote sensing and terrain are categorized as two data types, while they were combined as one in the text descriptions. Also, Field investigation data was not presented in Table 1.
Response: Thank you for the suggestion. We have reproduced Table 1 and added field investigation data to be consistent with the text. In the text, the remote sensing data and terrain data are described separately though they are both derived from remotely sensed data. The sources and uses of different data types are clearly displayed in revision.
- Regarding the sub-section “3.4. Land use pattern of 2025”, the result of this predication analysis should be further elaborated by highlighting the trend of changes as well as rates. It is also suggested elaborating further Table 8 by adding comparison between the situation of years 2020 and 2025.
Response: Thank you for your valuable comments. We have reproduced Table 8 and replaced "proportion" with "change between 2020 and 2025" and "average annual change" to further analyze future land-use change trends. In addition, by comparing the situation between 2020 and 2025, we can better understand the land use pattern of different land types in 2025 and their change rates.
- As the study also aims to provide scientific support for spatial planning of territory in Hefei City, the specific policy implications for this city should also be discussed.
Response: Thank you for the suggestion. We have added a part in Conclusions, which expect that our study may play an active role in implementation of different policies such as “Division of the Ecological Function in Anhui”, “Integrated Urban Planning of Hefei”. Our study may provide pertinent advice in optimization of land use and resource allocation to support Hefei's spatial planning of territory.
- As the integrated model LMCM model excludes experts’ knowledge, authors should clarify why this dimension is not important and what are the limitations when involving experts’ knowledge in terms of land use/cover change analysis and prediction.
Response: Thank you for your question. In the process of predicting land use change, unlike other existing models, LMCM model harnesses the coefficients of logistic regression analysis as weights for different driving factors and this does not require any human intervention, neither expert knowledge to assign weights for drivers. This process is able to avoid the subjective weight assignment and hence may allow LUCC modeling to be achieved with more reliable results in a more objective way than other models.
- The in-text citation style needs to be carefully checked and adjusted, as the cited numbers are mixed with the texts, while some of the citations are missing.
Response: Thank you for your suggestion. We carefully checked the style of citation in the text and adjusted the citation style as required by the Land journal in the revision.
Author Response File: Author Response.docx
Reviewer 3 Report
The goal of this manuscript is to develop an integrated hybrid (LMCM) model to predict land use/cover (LUC) pattern. It is an interesting research. My major concern is how come the accuracy the models are so high and then we can trust the results. Since some of the journals apply the reproducibility policy, I encourage the authors to release the data and the models for readers to reproduce their results. I only list some of the comments and I think the manuscript needs more efforts before it is to be considered publishing in "Land" and recommend its Major revision.
1. Please check your cited reference links.
2. The explanations of the 4 models are too simple, it is not easy for the readers to follow.
3. Table 3: accuracy of the LUC classification is very high. If the LUC categories are listed as in Table 2, could you please show the confusion matrices of the classification results. From the classification results showing in Figure 8, it is more realistic.
4. What are the major reasons for water areas changed from 1240 km2 (2010), to 1296 km2 (2015), then to 1110 km2 (2020)? Then 951 km2 (2025)? I do not think the elevation, slope, and distances from the rivers and lakes are the major reasons.
5. The weights from Figure 6 are only for LMCM model? How about the weights in the MCM model? I did not see the AHP (Line 271 mentioned) results.
6. Do socioeconomic factors (L1 and L2) and constraining factors (L3 and L4) need to give them weights for the models?
Some minor points:
1. Table 6: the variable symbols have shown in Table 5 already, it does not need to show again.
2. Figure 4: the variable symbols have shown in Table 5 already, it does not need to show again. You can refer them from Table 5. Same as Figure 6.
3. Line 247: What happened to the reference link? Same as lines 249, 251, 258, 259, 268, 283, 315, 451, and 477.
Author Response
Response to the reviewer’s comments:
To Reviewer #3
- Please check your cited reference links.
Response: Thank you for your suggestion. We have carefully checked the cited reference links in revision.
- The explanations of the 4 models are too simple, it is not easy for the readers to follow.
Response: Thank you for your comments. Sincec the research focus of this paper is to predict the LUCC by constructing the LMCM model and comparing it with the other three existing models. Therefore, the construction process of the LMCM model is emphatically introduced, and we provide corresponding references for the other three models for readers' reference.
- Table 3: accuracy of the LUC classification is very high. If the LUC categories are listed as in Table 2, could you please show the confusion matrices of the classification results. From the classification results showing in Figure 8, it is more realistic.
Response: Thank you for your comments. The confusion matrices of the classification results of 2010, 2015 and 2020 are shown in the following attached text:
- What are the major reasons for water areas changed from 1240 km2 (2010), to 1296 km2 (2015), then to 1110 km2 (2020)? Then 951 km2 (2025)? I do not think the elevation, slope, and distances from the rivers and lakes are the major reasons.
Response: Thank you for raising this issue. From the confusion matrices, the land cover classification of 2010, 2015 and 2020 were well achieved. The difference in water surface lies in the fact that there were difference in rainfall. We calculated the pre-acquisition rainfall of 6 months from August to February and found that they were respectively 371.2mm, 509mm and 300.2mm in 2009-2010, 2014-2015 and 2019-2020. That is why waters shows an increase from 2010 to 2014, and then, a decrease from 2015 to 2020.
As for that of 2025, it is the projected surface of waters as a part of the latter is likely to be converted into built-up area in 2025.
- The weights from Figure 6 are only for LMCM model? How about the weights in the MCM model? I did not see the AHP (Line 271 mentioned) results.
Response: Thank you for raising this issue. The weights in figure 6 are applicable to the hybrid LMCM model, but it can also be used in the MCM model. In this paper, the LMCM model is optimized on the basis of MCM model. Because the weights of MCM model are obtained through AHP; however, the way to obtain the weights within AHP is subjective. For this reason, instead of using AHP, we employed the “Completely standardized logistic regression coefficients” as weights in the LMCM model.
Since the focus of the research is on the construction of the hybrid LMCM model, the AHP results in MCM model is not presented in the paper.
- Do socioeconomic factors (L1 and L2) and constraining factors (L3 and L4) need to give them weights for the models?
Response: Thank you for your comments. L1 represents the protected zone of farmlands and L2 is the protected area of the first-class water source, which are both constraining factors, restricting the areas to be converted into other land use in the observation years to keep their status quo. The constraining factor is assigned in Boolean way, i.e., setting either 0 or 1, where "0" represents the area that is prohibited from conversion and "1" represents the area that can be convertible and developed. It is hence not necessary to give L1, L2, L3 and L4 weights in the LMCM model.
Minor points:
- Table 6: the variable symbols have shown in Table 5 already, it does not need to show again.
Response: Thank you for the suggestion. We have modified Table 6 according to your suggestion.
- Figure 4: the variable symbols have shown in Table 5 already, it does not need to show again. You can refer them from Table 5. Same as Figure 6.
Response: Thank you for the suggestion. We have modified Figure 4 and Figure 6 to remove the variable symbols.
- Line 247: What happened to the reference link? Same as lines 249, 251, 258, 259, 268, 283, 315, 451, and 477.
Response: Thank you for the suggestion. We rearranged the reference link to ensure that they meet the requirements of Land journal.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
I think the authors have addressed most of the review comments and provided appropriate justifications to the questions. I have no further comment.
N/A
Author Response
Reply to Reviewer 2
I think the authors have addressed most of the review comments and provided appropriate justifications to the questions. I have no further comment.
Reply: Thank you very much for provided valuable suggestions during the review process, which has greatly improved my paper. Thank you again for your efforts and wish you a happy life.
Author Response File: Author Response.docx
Reviewer 3 Report
The manuscript has been much improved after the revision. however, I am still not been convinced by the authors’ replies. I still think the accuracy the models are so high, and I cannot access the confusion matrices provided by authors. The authors also did not reply to the reproducibility policy idea. It is also difficult to image the water areas can be varied more than 345 km2 (1296-951). I am still confused how the weights were obtained. It still said by Analytic Hierarchy Process (AHP) in line 281. I guess the decision will be made by the editor. I do not have further comments.
Author Response
Reply to Reviewer 3
The manuscript has been much improved after the revision. however, l am still not been convinced by the authors' replies. I still think the accuracy the models are so high, and I cannot access the confusion matrices provided by authors.
Reply: Thank you for your comments. We supplied once again the confusion matrices for your examination. Please check the Appendix at the end of the document “Reply to Reviewer”.
The authors also did not reply to the reproducibility policy idea.
Reply: We do understand the reproducibility policy and have made efforts to make the approach and procedure understandable and reproducible.
lt is also difficult to image the water areas can be varied more than 345km2 (1296-951).
Reply：We have checked this difference carefully in multitemporal remote sensing images. This difference did exist between images of 2015 and 2020. As was explained in our previous reply to your comments, this difference was due to the difference in rainfall of pre-acquisition period from August to February. As for that of 2025, 951km2 of waters, is completely a predicted one. In terms of land use change trend, a large part of waters may be converted into built-up area.
I am still confused how the weights were obtained.
Reply: Within the LMCM model, the weights were actually obtained from the multiple logistic regression (MLR) analysis but not from AHP. More concretely, they are the fully standardized regression coefficients of MLR model.
lt still said by Analytic Hierarchy Process (AHP) in line 281.
Reply: It is not a conflict. Here is an introduction of MCM model in which the weights are determined by AHP. But what we proposed is the new hybrid LMCM model. We do not use AHP for deriving weights but apply MLR analysis for this purpose.
Appendix 1 Confusion Matrices
- LUC mapping of 2010
Confusion Matrix: C:\Users\hyc520xxy\OneDrive\桌面\333
Overall Accuracy = (42745/44358) 96.3637%
Kappa Coefficient = 0.9329
Ground Truth (Pixels)
Class river1_v building1_v cropland1_v forest1_v rangeland1_v
Unclassified 0 0 0 0 0
river1 [Blue1 1356 13 6 30 0
building1 [Re 27 1386 7 27 5
cropland1 [Gr 69 4 15527 1365 12
forest1 [Gree 3 0 6 24450 0
rangeland1 [Y 0 0 9 15 14
bareland1 [Ma 0 0 0 0 0
Total 1455 1403 15555 25887 31
Ground Truth (Pixels)
Class bareland1_v Total
Unclassified 0 0
river1 [Blue1 0 1405
building1 [Re 5 1457
cropland1 [Gr 10 16987
forest1 [Gree 0 24459
rangeland1 [Y 0 38
bareland1 [Ma 12 12
Total 27 44358
Ground Truth (Percent)
Class river1_v building1_v cropland1_v forest1_v rangeland1_v
Unclassified 0.00 0.00 0.00 0.00 0.00
river1 [Blue1 93.20 0.93 0.04 0.12 0.00
building1 [Re 1.86 98.79 0.05 0.10 16.13
cropland1 [Gr 4.74 0.29 99.82 5.27 38.71
forest1 [Gree 0.21 0.00 0.04 94.45 0.00
rangeland1 [Y 0.00 0.00 0.06 0.06 45.16
bareland1 [Ma 0.00 0.00 0.00 0.00 0.00
Total 100.00 100.00 100.00 100.00 100.00
Ground Truth (Percent)
Class bareland1_v Total
Unclassified 0.00 0.00
river1 [Blue1 0.00 3.17
building1 [Re 18.52 3.28
cropland1 [Gr 37.04 38.30
forest1 [Gree 0.00 55.14
rangeland1 [Y 0.00 0.09
bareland1 [Ma 44.44 0.03
Total 100.00 100.00
Class Commission Omission Commission Omission
(Percent) (Percent) (Pixels) (Pixels)
river1 [Blue1 3.49 6.80 49/1405 99/1455
building1 [Re 4.87 1.21 71/1457 17/1403
cropland1 [Gr 8.59 0.18 1460/16987 28/15555
forest1 [Gree 0.04 5.55 9/24459 1437/25887
rangeland1 [Y 63.16 54.84 24/38 17/31
bareland1 [Ma 0.00 55.56 0/12 15/27
Class Prod. Acc. User Acc. Prod. Acc. User Acc.
(Percent) (Percent) (Pixels) (Pixels)
river1 [Blue1 93.20 96.51 1356/1455 1356/1405
building1 [Re 98.79 95.13 1386/1403 1386/1457
cropland1 [Gr 99.82 91.41 15527/15555 15527/16987
forest1 [Gree 94.45 99.96 24450/25887 24450/24459
rangeland1 [Y 45.16 36.84 14/31 14/38
bareland1 [Ma 44.44 100.00 12/27 12/12
- LUC Mapping of 2015
Confusion Matrix: C:\Users\hyc520xxy\OneDrive\桌面\33
Overall Accuracy = (50989/51684) 98.6553%
Kappa Coefficient = 0.9733
Ground Truth (Pixels)
Class river1_v building1_v cropland1_v forest1_v rangeland1_v
Unclassified 0 0 0 0 0
river1 [Blue] 678 11 0 1 0
building1 [Re 8 3847 117 212 17
cropland1 [Gr 0 5 12879 112 47
forest1 [Gree 0 0 0 33551 0
rangeland1 [Y 0 0 0 0 34
bareland1 [Ma 0 0 0 0 0
Total 686 3863 12996 33876 98
Ground Truth (Pixels)
Class bareland1_v Total
Unclassified 0 0
river1 [Blue] 0 690
building1 [Re 17 4218
cropland1 [Gr 148 13191
forest1 [Gree 0 33551
rangeland1 [Y 0 34
bareland1 [Ma 0 0
Total 165 51684
Ground Truth (Percent)
Class river1_v building1_v cropland1_v forest1_v rangeland1_v
Unclassified 0.00 0.00 0.00 0.00 0.00
river1 [Blue] 98.83 0.28 0.00 0.00 0.00
building1 [Re 1.17 99.59 0.90 0.63 17.35
cropland1 [Gr 0.00 0.13 99.10 0.33 47.96
forest1 [Gree 0.00 0.00 0.00 99.04 0.00
rangeland1 [Y 0.00 0.00 0.00 0.00 34.69
bareland1 [Ma 0.00 0.00 0.00 0.00 0.00
Total 100.00 100.00 100.00 100.00 100.00
Ground Truth (Percent)
Class bareland1_v Total
Unclassified 0.00 0.00
river1 [Blue] 0.00 1.34
building1 [Re 10.30 8.16
cropland1 [Gr 89.70 25.52
forest1 [Gree 0.00 64.92
rangeland1 [Y 0.00 0.07
bareland1 [Ma 0.00 0.00
Total 100.00 100.00
Class Commission Omission Commission Omission
(Percent) (Percent) (Pixels) (Pixels)
river1 [Blue] 1.74 1.17 12/690 8/686
building1 [Re 8.80 0.41 371/4218 16/3863
cropland1 [Gr 2.37 0.90 312/13191 117/12996
forest1 [Gree 0.00 0.96 0/33551 325/33876
rangeland1 [Y 0.00 65.31 0/34 64/98
bareland1 [Ma 0.00 100.00 0/0 165/165
Class Prod. Acc. User Acc. Prod. Acc. User Acc.
(Percent) (Percent) (Pixels) (Pixels)
river1 [Blue] 98.83 98.26 678/686 678/690
building1 [Re 99.59 91.20 3847/3863 3847/4218
cropland1 [Gr 99.10 97.63 12879/12996 12879/13191
forest1 [Gree 99.04 100.00 33551/33876 33551/33551
rangeland1 [Y 34.69 100.00 34/98 34/34
bareland1 [Ma 0.00 0.00 0/165 0/0
- LUC Mapping of 2020
Confusion Matrix: C:\Users\hyc520xxy\OneDrive\桌面\44
Overall Accuracy = (35217/36057) 97.6704%
Kappa Coefficient = 0.9566
Ground Truth (Pixels)
Class river1_v building1_v cropland1_v forest1_v rangeland1_v
Unclassified 0 0 0 0 0
river1 [Blue] 1335 0 0 15 0
building1 [Re 17 1469 6 24 39
cropland1 [Gr 560 5 10436 150 3
forest1 [Gree 3 0 2 21822 0
rangeland1 [Y 0 1 0 0 44
bareland1 [Ma 0 0 0 0 0
Total 1915 1475 10444 22011 86
Ground Truth (Pixels)
Class bareland1_v Total
Unclassified 0 0
river1 [Blue] 0 1350
building1 [Re 3 1558
cropland1 [Gr 12 11166
forest1 [Gree 0 21827
rangeland1 [Y 0 45
bareland1 [Ma 111 111
Total 126 36057
Ground Truth (Percent)
Class river1_v building1_v cropland1_v forest1_v rangeland1_v
Unclassified 0.00 0.00 0.00 0.00 0.00
river1 [Blue] 69.71 0.00 0.00 0.07 0.00
building1 [Re 0.89 99.59 0.06 0.11 45.35
cropland1 [Gr 29.24 0.34 99.92 0.68 3.49
forest1 [Gree 0.16 0.00 0.02 99.14 0.00
rangeland1 [Y 0.00 0.07 0.00 0.00 51.16
bareland1 [Ma 0.00 0.00 0.00 0.00 0.00
Total 100.00 100.00 100.00 100.00 100.00
Ground Truth (Percent)
Class bareland1_v Total
Unclassified 0.00 0.00
river1 [Blue] 0.00 3.74
building1 [Re 2.38 4.32
cropland1 [Gr 9.52 30.97
forest1 [Gree 0.00 60.53
rangeland1 [Y 0.00 0.12
bareland1 [Ma 88.10 0.31
Total 100.00 100.00
Class Commission Omission Commission Omission
(Percent) (Percent) (Pixels) (Pixels)
river1 [Blue] 1.11 30.29 15/1350 580/1915
building1 [Re 5.71 0.41 89/1558 6/1475
cropland1 [Gr 6.54 0.08 730/11166 8/10444
forest1 [Gree 0.02 0.86 5/21827 189/22011
rangeland1 [Y 2.22 48.84 1/45 42/86
bareland1 [Ma 0.00 11.90 0/111 15/126
Class Prod. Acc. User Acc. Prod. Acc. User Acc.
(Percent) (Percent) (Pixels) (Pixels)
river1 [Blue] 69.71 98.89 1335/1915 1335/1350
building1 [Re 99.59 94.29 1469/1475 1469/1558
cropland1 [Gr 99.92 93.46 10436/10444 10436/11166
forest1 [Gree 99.14 99.98 21822/22011 21822/21827
rangeland1 [Y 51.16 97.78 44/86 44/45
bareland1 [Ma 88.10 100.00 111/126 111/111
Author Response File: Author Response.docx
