Next Article in Journal
Do Regional Differences in Forest Distribution Affect Residents’ Preferences for Forest Ecosystem Services?
Previous Article in Journal
Formulation and Numerical Verification of a New Rheological Model for Creep Behavior of Tropical Wood Species Based on Modified Variable-Order Fractional Element
 
 
Article
Peer-Review Record

Forest Carbon Storage Dynamics and Influencing Factors in Southeastern Tibet: GEE and Machine Learning Analysis

Forests 2025, 16(5), 825; https://doi.org/10.3390/f16050825
by Qingwei Fan 1, Yutong Jiang 1, Yuebin Wang 1,* and Guangpeng Fan 2,*
Reviewer 1: Anonymous
Reviewer 2:
Forests 2025, 16(5), 825; https://doi.org/10.3390/f16050825
Submission received: 11 March 2025 / Revised: 11 May 2025 / Accepted: 12 May 2025 / Published: 15 May 2025
(This article belongs to the Section Forest Ecology and Management)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Some general comments:

 

Writing and what is included in each section and in which order should be improved.

 

Some information on the performance of your model(s) should be included in the abstract.

 

When describing the used datasets, instead of giving general information, it is better to tell which variable, in which units, spatial and temporal resolution, name of the product, source and reference. For example, in reflectance, also the bands and wavelengths; in radar data, the channels and frequencies, etc. Common information such as period of study can be mentioned only once at the beginning or the end of the section.

 

When talking about the used algorithms (RF, CART, etc.), try not to repeat information. For example, expose all common traits to decision trees, mention you are using all of them for regression, etc. only once. Also, since you are doing regression, consider only explaining the relevant information for regression. No need to explain about classification.

 

The definitions of the metrics used to evaluate the performance of your models might not be needed since they are widely used by the scientific community. However, if you want to include them, summarize the common information (what yhat, ybar, y, and n are) at the beginning or the end of the section. By the way, yhat lacks a subindex most of the times it appears.

 

In Results 4.1:

Indicate the units of the RMSE, MAE, MAPE, intercept.

R^2 values in figures and Table 2 do not coincide. If ones are R2 and others R, please homogeneize.

Why invert the axis in figures? If training set and validation set scatterplots had the same axis, they would be easier to compare. Also indicate the variable and its units in each case.

Figure 4 below presents a particular distribution of samples. They seem to be in kind of vertical groups. I.e. the same predicted value for different observed values. Why does it happen? And why only in CART?

Please, homogeneize the use of uppercase or lowercase r for coefficient of determination.

 

In Results 4.2:

indicate which model was used to build images in figure 7.

Figure 7 f: indicate units of AGB changes.

 

In Results 4.3: data in Table 3 would be easier to visualize in the form of barplot.

 

In Results 4.4:

If the conversion factor from biomass to carbon in 0.5, figure 8 might not be needed. All comments regarding figure 8 can be made based on Figure 7.

Description in lines 493-497 does not fit images in Figure 8. Variables showed there do not present discrete values (classification) but continuous ones.

 

In Results 4.5: data in Table 4 would be easier to visualize in a plot as time series by land use.

 

In Results 4.6:

Which is the importance score for the rest of the inputs? Please consider building new models with fewer inputs (the most important ones) and check their performance. It would be great if similar R^2 and RMSE were achieved with less inputs.

 

In discussion:

Comment the limitations of the methodology and compare the results with similar studies.

 

Content that could fit better in Discussion section:

Lines 486-491.

Lines 501-502: “Possible … disasters”.

Lines 504-506: “This suggests … carbon stocks”.

Lines 515-517: “which may … degradation”.

Lines 524-530: “The phenomenon … development in the region”.

 

 

Specific comments:

 

Line 107: specify the variable showed in the figure.

Lines 114&122: 80% or 90%?

Lines 116-117: how can you say it is “one of the locations with the highest annual precipitation globally”? There is plenty of places in the world with more than 1000 mm per year.

Lines 128-129: it seems one “forest” is not necessary.

Line 140: you say classification, but you use the algorithms for regression.

Lines 153-157: confusing sentence, please rephrase.

Line 167: what do you mean by “seasonal”? Please clarify.

Lines 171-172: explain what GRD is if it is relevant.

Lines 243-244: same information was written in the precedent sentence.

Line 258: “tricks” does not seem an appropriate word for a scientific document.

Line 259: “classes”? Are not you doing regression?

Lines 271-275: this information does not belong to this section.

Line 277: what do you mean by “substituted”? Please clarify.

Lines 308-310: the definition of RPD seems confusing. Please rephase to fit the definition illustrated by equation (4).

Lines 326-327: specify the conversion factor(s) and its/their units.

Lines 339-341: this information talks about the goals of the study. It should be placed in an appropriate section.

Lines 515: “green” -> “blue”?

Line 525: “green” -> “blue”?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents the results of a study of carbon stocks in the forests of Southeastern Tibet, which is an important ecological security barrier on the Tibetan Plateau. Tibetan forests are of great importance for maintaining the ecological security of the region and provide the population with environmentally friendly food. The authors used the results of remote sensing to analyze the data. They developed a dynamic model for monitoring carbon stocks to assess spatial and temporal changes in Tibetan forests in the period 2019-2023. This made it possible to assess the degree of climate impact on humans. The results showed that the total carbon stock increases, taking into account spatial and temporal changes. The study is interesting, but there are questions and comments on the manuscript.
1. The authors need to focus on the measurement error of the remote sensing method. Since in the conclusions the authors cite changes in indicators over time from 75.46 Mg / ha in the first quarter of 2019 to 76.64 Mg / ha in the fourth quarter of 2023. Perhaps the variation of these values ​​​​is a measurement error. 2. In Section 3, the authors describe the methods used in too much detail. The text in this section can be shortened, but screenshots of the cartographic material on probing with these methods can be added.
3. Table 1 shows the methods and resolution from 10 to 20,000 meters. It is not clear what resolution is the minimum required to obtain high-quality results.
4. Section 4.1 should clearly describe what data is being discussed. What kind of data is in Table 2, what units of measurement. The text should be understandable not only to the authors, but also to the readers.
5. Figures 3-5, the names of the axes are unclear, and there are no units of measurement. What kind of data is this?
6. The results in Table 3 are unclear. How can carbon increase over the course of a year, then decrease, then increase again (the last three lines). Where does the carbon go in July-September 2023 and where does it come from later?
7. Table 4. What units of measurement and why are there two digits after the decimal point?
8. The list of references contains authors predominantly from one country. It is recommended to reasonably revise the list of references.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop