Determining the Accuracy of Structural Parameters Measured from LiDAR Images in Lowland Oak Forests

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

The paper is still interesting, however, the authors did not follow my suggestions, thus, the paper still needs revision:

1^st, they authors should not tell me, but the readers, how the age was determined. From my experience I know that the age in the management plans is not more than a poor guess, not based on any measurement.

2^nd, Maybe that the data for the full inventory are not publicly available, but at least the means of the investigated plots, with density, volume per hectare, etc. should give a rough insight what kind of stands were investigated.

In the new Figures 16 etc. the red line should be deleted, obviously there is no reason to fit a normal distribution, which is the assumption for t-tests.

Other than said in the answer to my request, nowhere the T was changed against the t in t-test

Comments on the Quality of English Language

There are minor changes necessary

Author Response

Comments 1:

1^st, they authors should not tell me, but the readers, how the age was determined. From my experience I know that the age in the management plans is not more than a poor guess, not based on any measurement.

• The aim of this research was to determine the accuracy of measuring parameters (height, breast diameter and volume) in the field and on LiDAR. For this, we did not need to know the exact age of the stand. It was enough for us to take it from the management program that is prepared for a particular area every 10 years.

 

Comments 2:

2^nd, Maybe that the data for the full inventory are not publicly available, but at least the means of the investigated plots, with density, volume per hectare, etc. should give a rough insight what kind of stands were investigated.

• Data on forests in the study area are available from the management program (on request) on the intranet of the company that manages state forests, and were not needed for this study. To achieve the research goal, a survey was conducted on the same plots in the field and on LiDAR. After statistical processing of the data (field - LiDAR), it was determined that there was no statistically significant difference and this was our input for automatic survey (on LiDAR) on larger areas. In the example for the Age class IV, the field measured data (part) are visible in Figure 9 (columns 11 and 13), as well as the data measured on LiDAR (columns 12 and 14).

 

Comments 3:

In the new Figures 16 etc. the red line should be deleted, obviously there is no reason to fit a normal distribution, which is the assumption for t-tests.

• Thank you for your suggestion, but other reviewers requested that the histograms remain in their original form.

 

Comments 4:

Other than said in the answer to my request, nowhere the T was changed against the t in t-test

• Sorry for the misunderstanding. It has now been corrected.

Reviewer 2 Report (Previous Reviewer 4)

Comments and Suggestions for Authors

The quality of figure is important. Please check all figure, especially for Figure 19.

Author Response

Comments 1:

The quality of figure is important. Please check all figure, especially for Figure 19.

• Thank you for your suggestion. The quality of all figures has been checked and Figure 19 has been corrected.

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

Mainly I would like to focus on the methodology, where I believe it needs more work before we can move on with the rest of the paper.

Find my comments in the pdf file and good luck.

 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

There are some parts where syntax and grammar need to be corrected. In some cases, there are phrases which are not clear. Have a closer look by reading the manuscript again and again and try to fix the "problematic" parts.

Author Response

Comments 1:

Add the location of the plots in fig.2 to fig 1. Figure 2 is unnecessary.

• Thanks for the suggestion, but we believe that both Figure 1 and Figure 2 should remain. Figure 1 shows only the research area, while Figure 2 shows only the areas where the research was conducted (which would be unclear if they were inserted into Figure 1).

Comments 2:

Figure 2 does not provide information about the age classes. The way you mention it in here requires additional info of age classes in fig.2 (to visually illustrate this piece of information).

• You are correct. The image is listed in the wrong place in the text. It has been corrected (highlighted in red on line 104).

 

Comments 3:

Add reference

• Reference has been added in line 122 and 575 (highlighted in red).

 

Comments 4:

The ground truth data were used from the FMP or you conducted this measurements?

• We conducted field measurements.

Comments 5:

Which one? Also, from the type and density of the forest as shown in the figures below I would like to know how did you ensure that tree heights were measured correctly (biased...you didnt miss the tree top...because of nearby branches etc). What was the distanc that you kept from each tree while measurign each tree and if you used the average from all 3 measurements during the measurement.

• In line 131-132, it was added with which instrument the heights were measured (highlighted in red).
• The height measurements were carried out exactly according to the rules prescribed for measuring the height of trees in the field. Also, the heights are determined in the winter period, when there are no leaves and the top of an individual tree can be clearly determined.

 

 

Comments 6:

You are using photogrammetric data but in your title you are mainly focusing on lidar. Why is that?

• We used photogrammetric data - an RGB camera in order to identify tree species.

 

Comments 7:

You need to add more details. This is very brief explanation.

• Line: 151 – 154: Here, only the standard data processing procedure in the used software is briefly described. In the rest of the text, the steps in data processing are explained in more detail and supported by images.

 

Comments 8:

For all those steps I would strongly recommend to read and use the following paper. https://www.sciencedirect.com/science/article/pii/S1569843222001182 Use this paper also to discuss your results.

• Thank you for your suggestion (line 428, 582-583).

 

 

Comments 9:

Removal of outliers is usually a pre-processing phase before your model is ready for further estimations.

• Thank you for your suggestion. It has been corrected in the text.

 

Comments 10:

You mainly focused on the mannual method of point removal especially in the lower sections of the tree (close to the ground) in order to be able to exract the dbh. The question is how practical that is when you have so many plots. Whats the practical aspect?

• Since the goal was to determine the accuracy of the field and LiDAR measurements, only on individual plots were points removed manually. Later, on larger areas (1 ha or larger), this was done automatically (as described in the paper).

 

Comments 11:

How did yoi manage to segment shrubs automatically? What was the error?

• In the tree classification process, a class of a certain height was selected, thus automatically eliminating shrubs.

 

Comments 12:

Automatic segmentation. Also, you need to explain that each tree color represents a different tree. Add a scale bar in this figure and the rest above.

• Thank you for your suggestion. The description of the Figure has been changed (line 189).

Comments 13:

Comments on the Quality of English Language

There are some parts where syntax and grammar need to be corrected. In some cases, there are phrases which are not clear.

• Proofreading of the entire article was done

Round 2

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

The authors have addressed all of my comments, and I am pleased with the overall presentation. I believe the paper is ready for acceptance in its current form."

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

Comments and Suggestions for Authors

I started with a good mood reviewing this manuscript, since the title and abstract are promising. However, at the end I can only conclude that the paper is not close to being publishable in a scientific journal. That is a pity, since a lot of effort and energy has been put in the data collection and analysis, so I will give some suggestions for improvement. However, they will be quite general.

Introduction:
You describe what research has been done in the field of LiDAR and forest inventories (although quite some extra literature has come out in the last year), but only describe what is possible. The knowledge gap is lacking, and as such also the need for this research is unjustified. What does your research add to this field?

Methods:
This section has major flaws, which makes it impossible to guarantee reproducibility by peers. Reproducibility is essential to safeguard the credibility of science.
* What type of scanner and camera did you use? What were the settings? Flight altitude? Patterns? You need to give more information on this.
* Why was the RGB data mentioned, since you don’t use it later.
* The LiDAR data processing is described as a black box, a piece of software which magically produces tree information. Which algorithms are behind it? Which settings did you use?
* The maps are too simple and lack essential information (e.g. figure 1 lacks a scale bar, geographical grid/location, north arrows, legend)
* It is very unclear how many plots were sampled and how many trees you have looked at in total.

Results:
The results are confusing and only analyzed at a shallow level. There is so much more information in this dataset, so please put more effort in analyzing this properly.
* My major issue is that I get lost when you talk about the number of plots or trees analyzed, for example in table 2 and table 4. Why are they different for d, v and h? You haven’t mentioned this anywhere in the methods section, so I assume every tree (or is it a plot?) can be analyzed for all those parameters, unless you decided to leave some out of the analysis. IF you did that you have to mention it somewhere and have very good reasons for it.
* Looking at e.g. figure 5 and the maps I assumed we would be looking at statistics of 1000’s of trees, but where do I see that in the results section? You got me totally lost on this.
* What is the use of Figure 6?
* Why do you decide to show only the results of a few age classes and not all of them? Make sure your results are complete and don’t cherry-pick them. Be complete.
* When looking at the tree height – where deviations are much larger – the description become really short. There must be more to tell about this! What do you trust more, your field measurements or the LiDAR data?

 

Discussion:
The discussion is basically lacking. It is very minimalistic and contains not comparison with existing literature of previous research. Unacceptable.

In line 297/298 it is mentioned that DBH measurements were not possible for thin trees. Where do I see this in your results? Please show data/figures that support this observation.

Smaller remark:
LiDAR doesn’t produce imagery, but pointclouds. Please don’t talk about LiDAR imagery.

 

Sorry for being so harsh, and maybe I misunderstood some parts, but please do a better job explaining what you did then. The way it is presented now just raises too many questions.

Author Response

Comments 1:

Introduction:
You describe what research has been done in the field of LiDAR and forest inventories (although quite some extra literature has come out in the last year), but only describe what is possible. The knowledge gap is lacking, and as such also the need for this research is unjustified. What does your research add to this field?

• We have added more recent literature to the introductory chapter (highlighted in red in the Literature section) and, in addition to the previously mentioned points, further emphasized the need for this research (lines 73–81).

Comments 2:

Methods:
This section has major flaws, which makes it impossible to guarantee reproducibility by peers. Reproducibility is essential to safeguard the credibility of science.
* What type of scanner and camera did you use? What were the settings? Flight altitude? Patterns? You need to give more information on this.

• Thank you for your suggestion. We have added all the requested information to the article text – lines 145–148 (highlighted in red).

* Why was the RGB data mentioned, since you don’t use it later.

• We used an RGB camera in order to identify tree species.

* The LiDAR data processing is described as a black box, a piece of software which magically produces tree information.

• Explanations have been added to the chapter – lines 157–174 (highlighted in red) along with images (5, 6, and 7).

* The maps are too simple and lack essential information (e.g. figure 1 lacks a scale bar, geographical grid/location, north arrows, legend)

• Figure 1 represents a schematic depiction of the study area (overview of the area), while Figures 2 and 3 have been revised.

* It is very unclear how many plots were sampled and how many trees you have looked at in total.

• Line 235 contains the number of plots and trees measured in the field and in the images. The condition was: “Within the systematic sampling grid (100x100 m), commonly used in field meas-ure-ments, and considering the variability, area, and age of the stand, the smallest ac-ceptable sample plot will be tested through random selection. This will ensure the achievement of equal measurement accuracy, ultimately contributing to a reduction in the time and costs of forest inventory planning.”
  
  

Comments 3:
Results:
The results are confusing and only analyzed at a shallow level. There is so much more information in this dataset, so please put more effort in analyzing this properly.
* My major issue is that I get lost when you talk about the number of plots or trees analyzed, for example in table 2 and table 4. Why are they different for d, v and h? You haven’t mentioned this anywhere in the methods section, so I assume every tree (or is it a plot?) can be analyzed for all those parameters, unless you decided to leave some out of the analysis. IF you did that you have to mention it somewhere and have very good reasons for it.

• For the comparison of field measurements and LiDAR measurements, the data (tree species, height, diameter at breast height, volume) were grouped into four age classes. Each age class contained a different number of trees that could be measured both in the field and using LiDAR (only trees for which both height and diameter at breast height could be measured were considered).

* Why do you decide to show only the results of a few age classes and not all of them? Make sure your results are complete and don’t cherry-pick them. Be complete.

• In the Results chapter, new images have been added to explain the results for the V. (Figures 13, 14, and 15) and VI. (Figures 16, 17, and 18) age classes. Additionally, results of automatic measurements using LiDAR have been included – randomly selected plots of 1 ha within the compartments (line 356 – 379).

*  What do you trust more, your field measurements or the LiDAR data?

• The higher measured heights in the field compared to the images are a result of insufficient distance from the tree during field measurements, which prevents a clear view of the tree's top. This leads to an error in height measurement, i.e., we overestimate the height, especially in stands with wider crowns. The reason is that, in order to improve the visibility of the top, the measurer stands too close, resulting in an overestimated height. Therefore, we place more trust in the results obtained from LiDAR.

Comments 4:

Discussion:
The discussion is basically lacking. It is very minimalistic and contains not comparison with existing literature of previous research. Unacceptable.

• We have expanded the Discussion section – lines 405–439, and added additional literature.

In line 297/298 it is mentioned that DBH measurements were not possible for thin trees. Where do I see this in your results? Please show data/figures that support this observation.

• Younger stands are heavily overgrown (dense), making it difficult to measure the diameter at breast height of thin trees. An example of a young stand is shown in the figure (in the appendix).

Smaller remark:
LiDAR doesn’t produce imagery, but pointclouds. Please don’t talk about LiDAR imagery.

• For data processing in the software, a point cloud was used; however, since it involved a combination of RGB imagery and LiDAR point clouds, the images of the clipped plots were formally referred to as images.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript deals with the accuracy of measuring individual tree measurements from Lidar images in lowland oak forests. This is an interesting paper, well prepared and well written, so that I have only a few remarks which could improve the work of the authors.

1.       From Figure 4 I see that the “plots” were circles, unfortunately I did not find the radius of the plots. Please supply.

2.       Please describe, how the age was determined, were trees cored?

3.       In order to better understand the structure of the investigated forests, I think that a table, giving the mean and the variation of the density (stem number per ha, basal area per hectare) mean height and volume per hectare and if possible the crown coverage of the investigated area and proportions of the species by stem number and volume.

4.       In all your tables: Please, give only a reasonable number of digits for the data. It makes no sense to give the volume by cubic millimeters (too many digits after the comma!). A usual rule is: 3 digits without leading and ending zeros.

5.       Please say, why the sample size for the whole area in Table 1 is different from the on in Table 2. The same is true for Table 3 and 4, 5 and 6 and 7 and 8.

6.       Please do not capitalize t-test. T-test would be understood as Tukey’s range test.    

7.       In Figure 4, please give the legend for the colors in English. I assume these are the species, but color is which species?

Author Response

1. From Figure 4 I see that the “plots” were circles, unfortunately I did not find the radius of the plots. Please supply.

• Thank you for your suggestion. The radius of the circle is provided in lines 102, 111 and 141 (highlighted in red).

2. Please describe, how the age was determined, were trees cored?

• All forests in Croatia are divided into management units, which are further subdivided into compartments and subcompartments. Each management unit has its own forest management plan, which is developed for a 10-year period and includes the age of the stands in each section. Since the research was conducted in lowland forests of pedunculate oak, the sections are divided by age classes (I. 1-20 years, II. 20-40 years, III. 40-60 years, IV. 60-80 years, V. 80-100 years, VI. 100-120 years, VII. 120-140 years), meaning all trees in a given section are of similar age. Therefore, no core samples were taken from the trees, and the age was obtained from the forest management plan.

3. In order to better understand the structure of the investigated forests, I think that a table, giving the mean and the variation of the density (stem number per ha, basal area per hectare) mean height and volume per hectare and if possible the crown coverage of the investigated area and proportions of the species by stem number and volume.

• The requested data are not publicly available.

4. In all your tables: Please, give only a reasonable number of digits for the data. It makes no sense to give the volume by cubic millimeters (too many digits after the comma!). A usual rule is: 3 digits without leading and ending zeros.

• Thank you for your suggestion – the data in the tables were calculated using the Statistica software.

5. Please say, why the sample size for the whole area in Table 1 is different from the on in Table 2. The same is true for Table 3 and 4, 5 and 6 and 7 and 8.

• For the comparison of field measurements and LiDAR measurements, the data (tree species, height, diameter at breast height, volume) were grouped into four age classes. Each age class contained a different number of trees that could be measured both in the field and using LiDAR (only trees for which both height and diameter at breast height could be measured were considered).

 

6. Please do not capitalize t-test. T-test would be understood as Tukey’s range test.

• Thank you for pointing out the error – it has been corrected!

7. In Figure 4, please give the legend for the colors in English. I assume these are the species, but color is which species?

• Thank you for pointing out the error – it has been corrected!

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

1. The paper needs to clarify many details about the experiment that was conducted.

2. Figures 7, 8, 9, and 10 need to be enlarged and explained more.

3. All abbreviations used in Tables 1, 3, 5, and 7 should be clarified.

4. Attention should be paid to and references should be modified, and more references should be used than those from 2023 and 2024.

5. There are additional notes that have been placed as comments within the text.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Need more

Author Response

1. The paper needs to clarify many details about the experiment that was conducted.

• Many details about the conducted research have been added throughout the chapters (highlighted in red).

2. Figures 7, 8, 9, and 10 need to be enlarged and explained more.

• Thank you for your suggestion.

3. All abbreviations used in Tables 1, 3, 5, and 7 should be clarified.

• Thank you for your suggestion.

4. Attention should be paid to and references should be modified, and more references should be used than those from 2023 and 2024.

• Recent literature has been added to the Literature chapter (highlighted in red).

5. There are additional notes that have been placed as comments within the text.

• Thank you for your suggestion.

 

Hoy many plots you used?

• Corrected in line 235.

Need more detailed about type of the processing?

• Explanations have been added to the chapter – lines 157–174 (highlighted in red) along with images (5, 6, and 7).

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

The title and abstract of this manuscript are quite appealing. While, there are so many problems, and I think they must be revised.

 

1.      For abstract.

There is no statistically significant difference, which is not a professional description. Please show the statistics.

2.      For introduction.

There only two paragraphs. Please segment and highlight the key point. At the same time, there is little research progress of Lidar.

3.      For innovation.

I think the innovation in the methods may be in line 149-164. The description of the part is not clear, I can not find the innovation.

4.      For discussion.

There are no references in this section, which I did not expect. This is unprofessional.

5.      For figure and table. Too many problems!

Figure 1. there are no three elements of the map; the color (grey/red/blue) is ugly; what is the gray big map? And can you replace the base map, like a vegetation map.

Figure 2. label the properties of the plots (line 92-95); what is in the bottom left corner?

Figure 3. What is the relationship with Figure 1 and Figure 2? I don't understand. What do the numbers 3a, 8a and so on stand for

Figure 4. I know you want to show your data frame, but this is very unprofessional. And please do not show ArcMap software directly.

Figure 6. there are two figure 6 (line 148/186). And line 186 do not need to display in the manuscript, please attached in the Supplementary material.

Last but not least, the quality of other figures and tables is also not good, please revise them all.

Author Response

1. For abstract.

There is no statistically significant difference, which is not a professional description. Please show the statistics.

• Thank you for your suggestion, but the statement that there are no statistically significant differences is supported by the research results and the statistical analyses of the data collected in the field and from LiDAR.

 

2. For introduction.

There only two paragraphs. Please segment and highlight the key point. At the same time, there is little research progress of Lidar.

• The Introduction chapter has been expanded (highlighted in red) and supplemented with recent literature.

 

3. For innovation.

I think the innovation in the methods may be in line 149-164. The description of the part is not clear, I can not find the innovation.

• The chapter has been supplemented and clarified.

 

4. For discussion.

There are no references in this section, which I did not expect. This is unprofessional.

• We have expanded the Discussion section – lines 405–439, and added additional literature.

 

 

5. For figure and table. Too many problems!

Figure 1. there are no three elements of the map; the color (grey/red/blue) is ugly; what is the gray big map? And can you replace the base map, like a vegetation map.

• The figure has been replaced.

 

Figure 2. label the properties of the plots (line 92-95); what is in the bottom left corner?

• The figure has been corrected.

 

Figure 3. What do the numbers 3a, 8a and so on stand for

• All forests in Croatia are divided into management units, which are further subdivided into compartments and subcompartments. The indicated numbers represent the subcompartments.

 

Figure 4. I know you want to show your data frame, but this is very unprofessional. And please do not show ArcMap software directly.

• The figure has been corrected.

 

 

Figure 6. there are two figure 6 (line 148/186). And line 186 do not need to display in the manuscript, please attached in the Supplementary material.

• Thank you for your suggestion.

 

Last but not least, the quality of other figures and tables is also not good, please revise them all.

• All figures have been corrected.

Author Response File: Author Response.docx