Modeling Height–Diameter Relationship Using Artificial Neural Networks for Durango Pine (Pinus durangensis Martínez) Species in Mexico

Round 1

Reviewer 1 Report

The paper uses different methods to simulate the breast height-height relationship of Durango Pine , which is an essential tool in forest management and planning. I believe that the manuscript needs to be revised in the following aspects.

1.The abstract section needs to be simplified, the purpose and main conclusions of the article need to be strengthened, and the description of the method section needs to be simplified.

2.The introduction section needs improvement. Firstly, the description of the research method in the preface needs to be simplified or deleted, as the second section provides a detailed introduction. Secondly, traditional methods have been widely applied to the relationship between tree height and diameter at breast height, and have achieved results. It is necessary to explain why the ANN method is still used and what problems exist with existing methods. Finally, the purpose and significance of this study should be further explained.

3.Some of the data in Table 1 are difficult to understand. The research data is applied to temporary forest inventory plots, where N represents the number of plants per hectare. Why is the minimum value 1? Does it mean that there is only one plant per hectare? Further sample information needs to be provided in this section.

Author Response

Modeling Height—Diameter Relationship Using Artificial Neural Networks for Durango Pine species in Mexico

Response to Reviewers:

Reviewer 1.

The paper uses different methods to simulate the breast height-height relationship of Durango Pine , which is an essential tool in forest management and planning. I believe that the manuscript needs to be revised in the following aspects.

R. Thanks for your comments and suggestions.

1.The abstract section needs to be simplified, the purpose and main conclusions of the article need to be strengthened, and the description of the method section needs to be simplified.

R. Thanks for your suggestions. The Abstract was rewritten and simplified.

2.The introduction section needs improvement. Firstly, the description of the research method in the preface needs to be simplified or deleted, as the second section provides a detailed introduction. Secondly, traditional methods have been widely applied to the relationship between tree height and diameter at breast height, and have achieved results. It is necessary to explain why the ANN method is still used and what problems exist with existing methods. Finally, the purpose and significance of this study should be further explained.

R. Thanks for your suggestions. The description of the research method in the preface was simplified and some references were deleted.

R. Thanks for your recommendations. An explanation about ANNs was added in the Introduction section. Also, we justify that ANNs had been used modeling this relationship in Mexican Forestry.

R. Thanks again for your comments. We appreciate it. The purpose of the study was explained and focused on in the last part of the introduction section.

3.Some of the data in Table 1 are difficult to understand. The research data is applied to temporary forest inventory plots, where N represents the number of plants per hectare. Why is the minimum value 1? Does it mean that there is only one plant per hectare? Further sample information needs to be provided in this section.

R. Thanks for your comments. We are so sorry. We made a mistake in the information of Table 1. The information about N and BA variables was updated in the revised version of Manuscript. The variables were changed. The minimum value of N is 10 trees per hectare and maximum 570 trees per hectare. In all cases the size of the plots was 0.10 ha.

Thank you so much for your comments and suggestions, we really appreciate it.

The revised version of Manuscript include the comments and suggestions.

Thanks again.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments for author File: Comments.pdf

Author Response

Modeling Height—Diameter Relationship Using Artificial Neural Networks for Durango Pine species in Mexico

Response to Reviewers:

Reviewer 2.

This manuscript explores the height-diameter (h-dbh) relationship of Durango Pine species in Mexico, employing the nonlinear mixed-effect modeling (NLMEM) and resilient backpropagation artificial neural network (RBPANN) methods. The results demonstrate superior performance of RBPANN over NLMEM in both training and testing phases, with RBPANN-tanh showing the best results. The study provides valuable insights for forest management and planning in Mexico.

R. Thank you so much for your comments. Thanks for reviewing and feedback to improve the manuscript.

Overall, the manuscript presents a promising contribution with practical significance. However, there are some minor suggestions and details that require attention:

R. Thanks.

In the abstract, provide a clear statement of the study's objectives. Consider adding a concise summary of the primary research goals. In the practical significance section, elaborate on the study's implications for Mexican forest management and planning.

R. Thanks. The abstract section was rewritten and organized.

In the introduction, offer a more explicit explanation of artificial intelligence's application and relevance in forest research, particularly the role of artificial neural networks (ANN) and resilient backpropagation artificial neural network (RBPANN) in forest modeling. Provide additional background information on the Durango Pine species, including its significance and characteristics within Mexican forests.

R. Thanks. Some sentences and paragraphs were rewritten in the Introduction section. Information about ecological and wood production for Durango pine was added.

Elaborate on the rationale for using unsupervised clustering analysis to group the dataset. Does it enhance model fitting? Have similar studies been conducted in different regions or for different tree species, and how do their findings compare to this study? Support these points with relevant references.

R. Thanks for comments. We add two paragraphs to explain more the unsupervised Clustering Analysis. Unfortunately, there are not similar studies to contrast the information. Thanks.

In the study area introduction, include geographic location, climate characteristics, vegetation composition, etc. In the experimental design and method steps, discuss potential sources of bias or errors that were considered.

R. Thanks. Done. Comments were included in the study area and the species composition included some paragraphs before. Thanks for comments.

Offer more comprehensive explanations for the figures and tables, including specific numerical values or result interpretations.

R. Thanks. Done.

In the discussion section, aside from Durango Pine, explore the general applicability of this method to other tree species and forest management scenarios. Discuss the potential for generalization to other forest regions.

R. Thanks. A paragraph was added in the discussion section.

Strengthen the conclusion section by providing a more specific summary of the research's main contributions and innovations, emphasizing its practical implications.

R. Thanks. We made it.

In conclusion, this manuscript requires minor revisions to address the mentioned points.

R. Thank you so much. We really appreciate it.

Author Response File: Author Response.pdf

Reviewer 3 Report

General remarks of the reviewer

The article presents well a proposal of using artificial neural networks to modeling the height-diameter at breast height relationship. It's an innovative tool for forest modeling. The correct sequence of content regarding the conducted research is maintained. After a analysis of selected publications, the research objectives are articulated in the final part of the introduction.

The following chapters requires some clarification:

Materials and Methods:

2.1. Study Area

In text and Figure 1, add GIS coordinates.

2.2. Dataset description

Please enter the size of inventory plots.

Add a regression function.

Table 1.

Please confirm that the density of Pinus durangensis Martínez ranged from 1 to 57 trees. This seems to be a very small share of the main specie.

Logic error: QMD-Max must be greater than Dm-Max.

Table 2.

Insert the correct title:Descriptive statistics for both training and testing dataset {?)

Variable: Error, change of h and dbh values.

 

3. Results

3.1. Training phase

3.1.1. NLMEM

Correct the mistakes in the commentary to table 4.

4. Discussion

Interestingly, however, it should be noted that the best model in the learning and testing phase (RBPANN-tanh) may not be fully satisfactory in practice when estimating the total tree height of Durango pine species. Based on the analysis of tables 1 and 2, we find a high variability of height (Hm=31%, h=43%), is the residuals ranged between -6 m and 6 m found on the basis of Figures 5 and 6 satisfactory? Determining the height as precisely as possible is of paramount importance for modeling the growth of the stand, in the first phase - together with age - determining the site index (bonitation), in the second as a functional feature for determining the volume.

 

5. Conclusions

 

The conclusions are constructive, please supplement with the need to continue research in the context of more effective use of artificial intelligence in forest modelling.

 

Technical Notes

Make the Figures  clearer if possible.

Eliminate logical errors and typos.

The description of the literature item needs to be corrected as required by the publisher: articles, books and other sources - italics of journal titles, year in bold, correct pages of journals and the access link and date of access in English. According to MDPI standard.

Details in the attached manuscript.

Comments for author File: Comments.pdf

Minor editing of English language required.

Author Response

Modeling Height—Diameter Relationship Using Artificial Neural Networks for Durango Pine species in Mexico

Response to Reviewers:

Reviewer 3.

The article presents well a proposal of using artificial neural networks to modeling the height-diameter at breast height relationship. It's an innovative tool for forest modeling. The correct sequence of content regarding the conducted research is maintained. After a analysis of selected publications, the research objectives are articulated in the final part of the introduction.

R. Thank you so much. Thanks for reviewing and comments in the entire manuscript. We really appreciate it.

The following chapters requires some clarification:

Materials and Methods:

2.1. Study Area

In text and Figure 1, add GIS coordinates.

R. Thanks. We included a new Figure and extreme coordinates were added in text. Thanks.

2.2. Dataset description

Please enter the size of inventory plots.

R. Thanks. Done.

Add a regression function.

R. We complemented the idea. This is about the variance explained for ten clusters in clustering analysis.

Table 1.

Please confirm that the density of Pinus durangensis Martínez ranged from 1 to 57 trees. This seems to be a very small share of the main specie.

R. Variables in Table 1 was restructured. You are right, 10 trees per hectare as minimum and 570 trees per hectare as maximum. N and BA were changed. Thanks for comments. Only Durango pine was considered.

Logic error: QMD-Max must be greater than Dm-Max.

R. Thanks for comments. We corrected it.

Table 2.

Insert the correct title:Descriptive statistics for both training and testing dataset {?)

R. Thank you. We made a mistake. We are sorry.

Variable: Error, change of h and dbh values.

R. Thank you so much. We really appreciate it. Variables were changed. First dbh and second h.

1. Results

3.1. Training phase

R. Thanks. Done.

3.1.1. NLMEM

R. Thanks. Done.

Correct the mistakes in the commentary to table 4.

R. Thanks for comments. We are so sorry for the mistakes.

1. Discussion

Interestingly, however, it should be noted that the best model in the learning and testing phase (RBPANN-tanh) may not be fully satisfactory in practice when estimating the total tree height of Durango pine species. Based on the analysis of tables 1 and 2, we find a high variability of height (Hm=31%, h=43%), is the residuals ranged between -6 m and 6 m found on the basis of Figures 5 and 6 satisfactory? Determining the height as precisely as possible is of paramount importance for modeling the growth of the stand, in the first phase - together with age - determining the site index (bonitation), in the second as a functional feature for determining the volume.

R. Thank you so much for comments and suggestions. Interesting ideas, we will work ahead with different species and variables. Yeah, the residuals showed a high variability, but it is common in this type of dataset because there are trees with similar dbh but different heights. Maybe, the interesting of this study is to give a new choice to predictions of heights and using the ANNs.

1. Conclusions

The conclusions are constructive, please supplement with the need to continue research in the context of more effective use of artificial intelligence in forest modelling.

R. Thank you so much. We rewrite some conclusions.

Technical Notes

Make the Figures  clearer if possible.

R. Done.

Eliminate logical errors and typos.

R. Thanks. Done.

The description of the literature item needs to be corrected as required by the publisher: articles, books and other sources - italics of journal titles, year in bold, correct pages of journals and the access link and date of access in English. According to MDPI standard.

R. We use the Endnote software and the MDPI style. But we will revised again.

Details in the attached manuscript.

R. Thank you so much. We really appreciate it. Your comment and suggestion let us to improve the entire manuscript.

Author Response File: Author Response.pdf