Studying Tropical Dry Forests Secondary Succession (2005–2021) Using Two Different LiDAR Systems
Round 1
Reviewer 1 Report
Studying Tropical Dry Forests secondary succession (2005-2021) using two different LiDAR systems, remotesensing-2499718-peer-review-v1
It’s a quite well organized manuscript to describe different stages of tropical dry forest in Costa Rica. A 16-year LIDAR data surveying provided very well seasonal differences between 2005 and 2021 with several statistical approaches.
1. It's a big issue in Canada wildforest fires in summer over the province of Alberta, Ontario and Quebec as a result of climate change. In the section of Discussion, I would suggest authors providing any applications with the methodology in this manuscript can contribute to natural-resources risks management of Canadian boreal mixedwood. It’s very helpful for the readers and audience who are interested in similar approaches and research activities.
2. Move the link of starting “Base map source: ESRI Satellite ….” into the caption of Figure 1.
3. In Figure 2, it would be great to reorganize the flowchart about processing of LVIS and RIEGL databases. Check the Figure 1 in the following reference. The simple, understandable flowchart gives good guidelines to potential stakeholders and scientists in conservation authorities of other geographical regions.
- Luo, X. et al. YOLOD: A Target Detection Method for UAV Aerial Imagery. Remote Sens. 2022, 14, 3240. https://doi.org/10.3390/rs14143240
4. In the caption of Figure 4, Is it the *.h5 waveform from RIGEL (2021) ? Check it out.
5. 3.3.2 Waveform metrics and 3.3.3 Waveform metrics ? Change the subtitle or remove one of them.
Author Response
Response to Reviewer 1 Comments
Many thanks for your comments and suggestions. Please see the attached file.
Point 1: It's a big issue in Canada wildforest fires in summer over the province of Alberta, Ontario and Quebec as a result of climate change. In the section of Discussion, I would suggest authors providing any applications with the methodology in this manuscript can contribute to natural-resources risks management of Canadian boreal mixedwood. It’s very helpful for the readers and audience who are interested in similar approaches and research activities.
Response 1: Thank you for the comment. We definitely care about the ecosystem and want to make more contributions to society. However, our focus aim on the Tropical Dry Forests, which are not a type of ecosystem grew inner Canada. It is hard to make the relationship, and it will jeopardize the coherence of the introduction. We see very little applicability unless we could say that changes in tree high over specific period of time can be integrated into fire prediction/biomass models. We can probably add few lines on the discussion.
Point 2: Move the link of starting “Base map source: ESRI Satellite ….” into the caption of Figure 1.
Response 2: Thank you for the comment. It is already changed.
Point 3: In Figure 2, it would be great to reorganize the flowchart about processing of LVIS and RIEGL databases. Check the Figure 1 in the following reference. The simple, understandable flowchart gives good guidelines to potential stakeholders and scientists in conservation authorities of other geographical regions.
- Luo, X. et al. YOLOD: A Target Detection Method for UAV Aerial Imagery. Remote Sens. 2022, 14, 3240. https://doi.org/10.3390/rs14143240
Response 3: Thank you for the comment. We would care if the chart could directly illustrate the processing of our works. The chart is hard to maintain the regular flowchart with YES and NO decisions or steps followed by step. Moreover, it is the whole processing related, so using fewer lines and shapes to illustrate is hard.
Point 4. In the caption of Figure 4, Is it the *.h5 waveform from RIGEL (2021) ? Check it out.
Response 4: Thank you for the comment. It is correct, it means the RIEGL file had already converted to “.h5” file. From 183-200 we illustrate the processing.
Point 5. 3.3.2 Waveform metrics and 3.3.3 Waveform metrics ? Change the subtitle or remove one of them.
Response 5: Thank you for the comment. It is already changed. “3.3.3. Age group and succession stages”
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript titled “Studying Tropical Dry Forests secondary succession (2005-2021) using two different LiDAR systems” presents a study that uses two different airborne LiDAR systems to quantify structural changes in the forest at Santa Rosa National Park. The authors used line- and shape-based waveform metrics to record the overall changes in the TDF structure. The paper takes into a worthy topic, and I believe the document would make a nice contribution to the Remote Sensing community; however, I have some concerns before the manuscript goes to further processing.
Here are my revisions:
First, please check Lines 119-132 and 137-150 because they are repeated. Please modify accordingly.
From the methods, my major concern is the geographic correspondence between the RIEGL and LVIS data. How did the authors ensure the overlapping of the footprints of data?
Later, for the upscaling process, Figure 4 illustrates de procedure, but I was thinking about what happened to the natural “gaps” between the RIEGL footprint. Are the aggregation/upscaling process adding some sources of uncertainties? Furthermore, if so, are they accounted for or neglected?
Are section 3.3.2 the same as 3.3.3? They have the same name.
As the authors mentioned, the differences between the dry and wet seasons introduced some “noise” due to the number of leaves in every season. Therefore, the waveform metrics analysis does not describe or explain if a “stage” of succession is due to the lack of presence of leaves. Moreover, the 3.3.3 section is confusing. How did the authors identify the stages of succession described in lines 280 and 283? I think the important question is, at what stage of succession was the forest in 2005? The authors stated that because 16 years have passed between images, the age of the forest changed, but it does not mean that the structure changed too.
Later in the results section 4.1, “Change in relative height traits and canopy height” is introduced, but in the methods, there is no description of how canopy changes are measured, at least not explicitly.
The manuscript needs a deep revision, especially to structure and flow of the methods.
The quality of the English Language is not an issue.
Author Response
Response to Reviewer 2 Comments
Many thanks for your comments and suggestions. Please see the attached file.
Point 1: First, please check Lines 119-132 and 137-150 because they are repeated. Please modify accordingly.
Response 1: Thank you for the comment. It is already deleted.
Point 2: From the methods, my major concern is the geographic correspondence between the RIEGL and LVIS data. How did the authors ensure the overlapping of the footprints of data?
Later, for the upscaling process, Figure 4 illustrates de procedure, but I was thinking about what happened to the natural “gaps” between the RIEGL footprint. Are the aggregation/upscaling process adding some sources of uncertainties? Furthermore, if so, are they accounted for or neglected?
Response 2: Thank you for the comment. From ‘3.3 Pseudo-waveform synthesis (RIEGL to LVIS)’, we introduced the methods in how to use the geolocation of the LVIS as the ‘filter’ to select the small footprint inside also; from Figure 4, we use the image to show the footprint of the LVIS. And then, yes, we cannot avoid the gaps. However, our calculations aim to reduce noise and uncertainty since the waveforms are directly derived from the cloud points, so the gaps will not be calculated. Furthermore, here is the link to the “rGEDI” code: https://github.com/carlos-alberto-silva/rGEDI.
Moreover, for the raw data, we clean the data before using ‘rGEDI’ to produce the waveforms. At the same time, the waveforms will have some insufficient data, and they all were cleaned.
Point 3: Are section 3.3.2 the same as 3.3.3? They have the same name.
Response 3: Thank you for the comment. It is already changed. “3.3.3. Age group and succession stages”
Point 4. As the authors mentioned, the differences between the dry and wet seasons introduced some “noise” due to the number of leaves in every season. Therefore, the waveform metrics analysis does not describe or explain if a “stage” of succession is due to the lack of presence of leaves. Moreover, the 3.3.3 section is confusing. How did the authors identify the stages of succession described in lines 280 and 283? I think the important question is, at what stage of succession was the forest in 2005? The authors stated that because 16 years have passed between images, the age of the forest changed, but it does not mean that the structure changed too.
Response 4: Thank you for your comment. While many studies have discussed the uncertainty of forest stages, we are determined to find the ultimate answer. Based on Dr. Zhao's team's work, we can confidently locate different age groups and stages in LVIS 2005 data. The main issue is that we do not know the real-time stage area of the forest due to its dynamics, so we calculate the same stage areas with 16 years of changes. In the future, we hope to produce a 2021 stage map and age map to aid in further studies. Regarding which stage of the forest was in 2005, the random forests are mixed with 3 stages, and no specific stage could represent the whole forest. As we mentioned before, from Dr. Zhao's team's work, we can locate the area in 2005. However, we lacked the accurate age map in 2021, so the best way is to use the same area plus the 16 years to produce a new age map.
Point 5. Later in the results section 4.1, “Change in relative height traits and canopy height” is introduced, but in the methods, there is no description of how canopy changes are measured, at least not explicitly.
Response 5: Thank you for the comment. Actually, these changes are introduced in Table 2 and 3.3.1 Tree height from the waveform. Please have a look.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Dear authors:
Many thanks for carefully addressing my revisions. I am satisfied with the answers provided. As a final revision, I recommend including response 4 in the discussion section since I believe it provides a good point of discussion for the context of the study and future directions. Please consider it.
The quality of the English Language is a minor issue. Although, a final revision is necessary for style and grammar check.
Author Response
Dear reviewer, we thank you for the time taken to review our responses to your comments. Based on your assessment we have implemented the following two items:
- " I recommend including response 4 in the discussion section since I believe it provides a good point of discussion for the context of the study and future directions. Please consider it."
R:/ We have added a new paragraph at the start of the discussion to cover this issue. The new paragraph can be found on lines 472-477.
2. "The quality of the English Language is a minor issue. Although, a final revision is necessary for style and grammar check."
R:/ We have done a complete review of the manuscript using PaperPal to address this issue.
We hope that the changes are satisfactory.
