Mapping Forest Restoration Probability and Driving Archetypes Using a Bayesian Belief Network and SOM: Towards Karst Ecological Restoration in Guizhou, China
Round 1
Reviewer 1 Report
I wish to congratulate the authors for the interesting manuscript. Uncertainty in forest restoration programmes is a crucial issue. The proposal for a methodology mapping the spatial distribution of the forest restoration probability is of certain interest to the readers. Overall, the manuscript is well organized and the purposed method is sound.
The research covers the Guizhou Province in China by challenging issues from different aspects such as time series, self-organizing maps (SOM), and Bayesian networks. I am fully satisfied with how the authors established the entire experiment and discussed the results.
My comments are as follow:
- for section "2.1 Study area": since the karst landforms are a key issue, the soil map for the study case, showing the kast areas distribution might be useful to the readers
- line 142 to 156: among the databases collected to obtain the driving factors, I see elements were ranging from very different spatial resolutions here. For instance, the soil dataset from the national Tibetan Plateau data center at the scale of 1:1'000'000 (what spatial resolution?) is used together with the nighttime light and the NDVI with a spatial resolution of 0,5 and 1 km. Although I am ok that all datasets had been resampled to a 1km pixel size before being projected, the differences in data resolution should be better addressed.
- Line 172: the hexagonal plane prototypes configuration test as input dataset might be cryptic for some readers. Could the authors ease this critical passage with a more detailed description?
- Figure 2: The caption should end at the end of the sentence"...for forest restoration". The variables list can be better moved in the text body or a table.
that's all
Author Response
Dear reviewer,
First of all, we are very thankful for your constructive comments on our study. Specially, we are heartily grateful to your valuable suggestions.
The manuscript has been revised carefully and strictly according to the minor comments provided in your letter. Moreover, this manuscript was edited for proper English language, grammar, punctuation, spelling, and overall style by one or more of the highly qualified native English speaking editors at NativeEE. NativeEE specializes in editing and proofreading scientific manuscripts for submission to peer-reviewed journals.
We are submitting our revised version entitled “Mapping forest restoration probability and driving archetypes using Bayesian belief network and SOM: towards Karst ecological restoration in Guizhou, China”, Manuscript ID remotesensing-1568056. Please find the revised manuscript with track changes. In order to facilitate your review, Red highlight was used to show revision and changes. In the following “Point-by-point response to your comments”.
Please do not hesitate to contact me, if further material or information is needed.
Thanks again.
Sincerely yours,
Li Peng
General comment:
- I wish to congratulate the authors for the interesting manuscript. Uncertainty in forest restoration programmes is a crucial issue. The proposal for a methodology mapping the spatial distribution of the forest restoration probability is of certain interest to the readers. Overall, the manuscript is well organized and the purposed method is sound.
The research covers the Guizhou Province in China by challenging issues from different aspects such as time series, self-organizing maps (SOM), and Bayesian networks. I am fully satisfied with how the authors established the entire experiment and discussed the results.
Specific comments:
- for section "2.1 Study area": since the karst landforms are a key issue, the soil map for the study case, showing the kast areas distribution might be useful to the readers
Response: Thank you very much for your proposal. Following your suggestion, we have used the landforms map to display the uniqueness of the study area, as shown in Figure 1.
- line 142 to 156: among the databases collected to obtain the driving factors, I see elements were ranging from very different spatial resolutions here. For instance, the soil dataset from the national Tibetan Plateau data center at the scale of 1:1'000'000 (what spatial resolution?) is used together with the nighttime light and the NDVI with a spatial resolution of 0,5 and 1 km. Although I am ok that all datasets had been resampled to a 1km pixel size before being projected, the differences in data resolution should be better addressed.
Response: Thank you for your comments. I am very sorry for this trouble. To solve this problem, we have supplemented the data source and data processing process in line 143-145, line155-160 and line 163-168.
Line 143-145: In order to improve the image quality, we preprocessed the data via radiometric calibration, atmospheric correction, image mosaic and clipping, cloud removal, shadow processing and spectral normalization.
Line 155-160: The spatial grids of temperature, precipitation and potential evapotranspiration with 1-km resolution were interpolated using the meteorological interpolation software Anusplin. Soil data including soil texture and soil depth with a spatial resolution of 1 km were obtained from the 1:1,000,000 soil database of the National Tibetan Plateau Data Center (http://data.tpdc.ac.cn/).
Line 163-168：The distances of sample locations from roads and settlements were obtained using the near tool in ArcGIS10.6. The afforestation area data was determined with the help of the local governmental departments. The afforestation area of each grid was obtained via geostatistical analysis. In this study, all the selected driving factors were resampled into a raster with 1 km ´1 km pixel size to ensure the same spatial resolution and grid number.
- Line 172: the hexagonal plane prototypes configuration test as input dataset might be cryptic for some readers. Could the authors ease this critical passage with a more detailed description?
Response: Thank you for your comments. According to your comments, the concrete information of “hexagonal plane prototypes configuration test” has been supplemented.
Line182-198: In the second step, we completed the parameterization of SOM by defining a priori number of clusters in a two-dimensional plane. This step is crucial as defining too many clusters may lead to the separation of relatively homogeneous clusters, while defining too few clusters may lead to inhomogeneity with high variability of input data[15]. To define an appropriate number of clusters, we analyzed the number of hexagonal plane prototypes in different combinations (e.g., 5 × 5 vs. 10 × 10) based on the Davies–Bouldin (DB) index and the mean distance of samples in each cluster [27]. In this study, we selected a 5 × 5 hexagonal plane for the drivers associated with forest coverage change as the DB index (6.52) and the mean distance (7.02) to the cluster centroids was more satisfactory at this point. The SOM method was used to generate a monolayer map of the clusters of forest coverage change drivers. Finally, an actual monolayer map of the clusters was generated iteratively. The SOM method involves creation of patterns from factors based on similarities and differences [19]. The optimal clustering mode was obtained and codebook vectors were used to detect the relative importance of each factor under each archetype. This method facilitated the identification of the impact of spatial allocation of driving factors on forest coverage change.
- Figure 2: The caption should end at the end of the sentence"...for forest restoration". The variables list can be better moved in the text body or a table.
Response: Thank you for your comments. Following your comments, we have revised the caption of Figure 2.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript analyzes the ecological restoration in Guizhou province between 2005 and 2018. They have also shown the interaction between 16 driving factors within the region. This study is relevant to showcase the implementation of current restoration projects and their future importance towards improving local ecology. Although, I have some suggestions to the authors which I have listed below.
- There are several grammatical mistakes throughout the manuscript that needs to be corrected.
- There is no literature about why they used SOM and BBN for this study. Is there a possibility of any other method that would have given a better ROC?
- Line 49-51, The statement is not clear here, if possible then try to split it into two separate sentences.
- Line 59, what do you mean by "forest repair"? please specify
- Line 66-67, Again, here it is not clear about what "certain dynamic trends" authors are talking about. Explanation with an example might be useful here.
- Line 115-116, Again, here it is not clear about what "certain dynamic trends" authors are talking about. Explanation with an example might be useful here.
- Line 120-121, Please include a reference when you are providing a statistics about the area.
- Line 137-138, please rephrase the sentence
- Line 168, it should be "Z-score<0"
- Line 171, "technology" is not a right word to use here, you can use something like method or algorithm etc.
- Table 2, Please check the values below. For Example, The lowest range of Tm is 16-18, but isn't it should be <12? similartly for other parameters like Pm, slope etc.
- Figure 6 & 7, make them as a single figure. As they are showing the same thing.
Author Response
Dear reviewer,
First of all, we are very thankful for your constructive comments on our study. Specially, we are heartily grateful to your valuable suggestions.
The manuscript has been revised carefully and strictly according to the minor comments provided in your letter. Moreover, this manuscript was edited for proper English language, grammar, punctuation, spelling, and overall style by one or more of the highly qualified native English speaking editors at NativeEE. NativeEE specializes in editing and proofreading scientific manuscripts for submission to peer-reviewed journals.
We are submitting our revised version entitled “Mapping forest restoration probability and driving archetypes using Bayesian belief network and SOM: towards Karst ecological restoration in Guizhou, China”, Manuscript ID remotesensing-1568056. Please find the revised manuscript with track changes. In order to facilitate your review, Red highlight was used to show revision and changes. In the following “Point-by-point response to your comments”.
Please do not hesitate to contact me, if further material or information is needed.
Thanks again.
Sincerely yours,
Li Peng
General comment:
The manuscript analyzes the ecological restoration in Guizhou province between 2005 and 2018. They have also shown the interaction between 16 driving factors within the region. This study is relevant to showcase the implementation of current restoration projects and their future importance towards improving local ecology. Although, I have some suggestions to the authors which I have listed below.
Specific comments:
- There are several grammatical mistakes throughout the manuscript that needs to be corrected.
Response: Thank you for your suggestions. According to your advice, this manuscript was edited for proper English language, grammar, punctuation, spelling, and overall style by one or more of the highly qualified native English speaking editors at NativeEE. NativeEE specializes in editing and proofreading scientific manuscripts for submission to peer-reviewed journals.
- There is no literature about why they used SOM and BBN for this study. Is there a possibility of any other method that would have given a better ROC?
Response: Thank you for your proposal. We have supplemented a detailed description to explain why they used SOM and BBN for this study in line76-80 and line 89-97. In addition, these two methods have been used to solve the interaction of driving factors and uncertainty in forest restoration assessment, and the application of ROC in this paper is to prove the reliability of the methods. Other models may have better ROC values, but the selection of models should first focus on what problem it can solve, and then obtain the optimal model based on the iteration of the algorithm. Therefore, the ROC value of the method selected in this study may not be the largest, but it is the most suitable to solve the problem concerned in this study.
- Line 49-51, The statement is not clear here, if possible then try to split it into two separate sentences.
Response: Thank you for your comments. Following your comments, we have revised the sentence in line45-46.
Line45-46: However, the prediction of forest restoration requires insight into the spatial patterns and driving factors of forest systems to formulate reasonable restoration measures.
- Line 59, what do you mean by "forest repair"? please specify
Response: It is an important question. We thank you for your valuable comments, and all of the “forest repair” expressions have been revised to “forest restoration” throughout manuscript.
- Line 66-67, Again, here it is not clear about what "certain dynamic trends" authors are talking about. Explanation with an example might be useful here.
Response: Thank you for your careful review. Following your suggestion, we have adjusted the expression and given an example to explain it. The revisions can be found in line62-64.
Line62-64: Further, previous studies are limited in terms of modeling and data availability, and depict only quantitative changes and trends in forest coverage.
- Line 115-116, Again, here it is not clear about what "certain dynamic trends" authors are talking about. Explanation with an example might be useful here.
Response: Thank you for your comments. Following your suggestion, we have adjusted the expression.
- Line 120-121, Please include a reference when you are providing a statistics about the area.
Response: Thank you for your comments. In order to make the statistics sound enough, we quoted relevant articles to support it.
- Line 137-138, please rephrase the sentence
Response: Thank you for your comments. According to your comments, the expression has been corrected.
Line 142-144: In order to improve the image quality, we preprocessed the data via radiometric calibration, atmospheric correction, image mosaic and clipping, cloud removal, shadow processing and spectral normalization.
- Line 168, it should be "Z-score<0"
Response: Thank you for your careful review. The wrong was corrected.
- Line 171, "technology" is not a right word to use here, you can use something like method or algorithm etc.
Response: Thank you for your careful review. We have revised the expression.
- Table 2, Please check the values below. For Example, The lowest range of Tm is 16-18, but isn't it should be <12? similartly for other parameters like Pm, slope etc.
Response: Thank you for your comments. In this study, the driving factors were divided using FR models. First, we have ranked the driving factors based on the susceptibility of each attribute interval of the factor to the event. Then, the intervals with similar frequency ratios have been be merged to realize the scientific division of indicator factor status. Therefore, the “Lowest” status is not the lowest range.
- Figure 6 & 7, make them as a single figure. As they are showing the same thing.
Response: Thank you for your comments. According to your comments, we have combined the Figures 6 & 7.
