Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This work, titled “Detecting Woody Plant Cover in the Foothills Parkland and Montane Ecoregions of Southern Alberta”, aims to detect woody plant encroachment using high-resolution PlanetScope data. Overall, this work holds important scientific and practical value for grassland management and conservation. The suggestions are as follows:

(1) How is the “early stage” determined? How are such samples selected? What unique characteristics does this stage possess? And how does this project achieve effective detection or identification of this stage? (The paper earlier stated “... the early stages of WPE have proven particularly challenging to monitor”).

(2) The authors mention in Section 2.4 that “The high-resolution image selection is based on the need to detect woody cover in early stages (Lines 174-175).” This statement is not detailed enough, and readers may find it difficult to fully understand. It is recommended to provide supplementary explanations.

(3) Table 1 is crucial. Currently, the meaning of the values in this table is not intuitive and is difficult to understand; optimization is recommended. The same issue also exists in Table 5. Please revise them together.

(4) Judging from Figure 2, the fitting effect of the models is not ideal (the R² values are all relatively small). It is recommended to provide an in-depth explanation or discussion on this.

(5) It is recommended to improve the content regarding field data collection in Section 2.2. The quality of the sample data in this section directly determines the reliability of the subsequent analysis results.

(6) It is recommended to add an elaboration on the importance of the study area in “1. Introduction”.

 

Other minor issues:

(1) A space needs to be added between numbers and units (e.g., m, km, km²). Please check the entire manuscript and modify them uniformly.

(2) Figures 2, 5, and 6 all contain subplots (A, B, ...), but they are not detailed in the corresponding figure captions. Please supplement this.

(3) Please carefully check the expression of statistics throughout the text (including figures and tables) to ensure a standardized and unified format (e.g., R²/r², P/p, etc.).

Author Response

Comment 1: How is the “early stage” determined? How are such samples selected? What unique characteristics does this stage possess? And how does this project achieve effective detection or identification of this stage? (The paper earlier stated “... the early stages of WPE have proven particularly challenging to monitor”).

Response 1: A great comment. It is difficult to define “Early stage”. Previous studies (Soubry & Guo, 2022; Pu et al., 2025) used less than 25% woody cover as early detection. However, woody cover is a continuous variable, spanning from 0-100. There is some imprecision because there are not clearly defined breakpoints within that range. The “early stages” are largely characterized by their state of lower or less dense cover, which is primarily of interest due to that state having the greatest likelihood of responding to management-based intervention (Fogarty et al., 2025). This project, to some degree, achieves the identification of that state with the use of relatively novel visual bands and vegetation indices.

To make the above points clearer more information has been added to the abstract (Page 2, Line 37), and Introduction (Page 3, Line 85).

Pu, Y., Lu, X., Soubry, I., & Guo, X. (2025). Early detection of woody plant encroachment in Canadian prairies using UAV imagery and transformer-based deep learning. Ecological Informatics, 90. https://doi.org/10.1016/j.ecoinf.2025.103354

Soubry, I., & Guo, X. (2022). Invasive and native woody plant encroachment: Definitions and debates. Journal of Plant Science and Phytopathology, 6(2), 084–086. https://doi.org/10.29328/journal.jpsp.1001079

Comment 2: The authors mention in Section 2.4 that “The high-resolution image selection is based on the need to detect woody cover in early stages (Lines 174-175).” This statement is not detailed enough, and readers may find it difficult to fully understand. It is recommended to provide supplementary explanations.

Response 2: Thank you for the feedback. The phrasing of Section 2.4 (Page 7, Lines 179-180) was altered to make the intention of the choice of imagery clearer.

Comment 3: Table 1 is crucial. Currently, the meaning of the values in this table is not intuitive and is difficult to understand; optimization is recommended. The same issue also exists in Table 5. Please revise them together.

Response 3: Thank you for this feedback. Tables have all increased in number due to the inclusion of a new summary table. The caption of Table 2 (formerly table 1) was adjusted to better emphasize the contents of the table (Page 6). Table 6 (formerly Table 5) was also reformatted to better convey the results (Page 11).

Comment 4: Judging from Figure 2, the fitting effect of the models is not ideal (the R² values are all relatively small). It is recommended to provide an in-depth explanation or discussion on this.

Response 4: Thank you for your feedback. Due to this comment in conjunction with the recommendation of Reviewer 2, the figure has been removed. We agree that further discussion of the contents of that table would still improve the manuscript quality. Further reference to this has been added to Section 4.1 (Page 14, Lines 334-336). We elaborate more on the natural complexity and heterogeneity in the plant community that contributes to those results.

Comment 5: It is recommended to improve the content regarding field data collection in Section 2.2. The quality of the sample data in this section directly determines the reliability of the subsequent analysis results.

Response 5: Thank you for the suggestion. Together with comment from Reviewer 3, additional information was added to (Page 4, Lines 143-145, 147-149).

Comment 6: It is recommended to add an elaboration on the importance of the study area in “1. Introduction”.

Response 6: Thank you for the suggestions. Addition information was added about the importance of the study area (Page 2, Lines 63-64).

Comment 7: A space needs to be added between numbers and units (e.g., m, km, km²). Please check the entire manuscript and modify them uniformly.

Response 7: Corrected.

Comment 8: Figures 2, 5, and 6 all contain subplots (A, B, ...), but they are not detailed in the corresponding figure captions. Please supplement this.

Response 8: That you for this comment. The figure captions were updated to include the subplots.

Comment 9: Please carefully check the expression of statistics throughout the text (including figures and tables) to ensure a standardized and unified format (e.g., R²/r², P/p, etc.).

Response 9: Thank you for this comment. The expression of statistics has been standardized throughout the text.

Reviewer 2 Report

Comments and Suggestions for Authors

Your work shows a careful fieldwork, with matching dates from the satellite pass, which is good. The results are interesting too.  I will suggest some minor adjustments to improve the reading.

Did you test your data for normality, right?

What the lines inside the study area ranches mean? Are they different "farms"?

I don't think Figure 2 is needed, as you comment the results in the text. You can present the graphs in the Appendix.

I did not understand Table 5, what the data mean. Please clarify.

Figure 3 presents the results of the best model you showed earlier? The one with red edge and yellow bands? Please clarify that.

In line 278 please add r2 instead of r (for consistency with previous results).

 

Author Response

Comment 1: Did you test your data for normality, right?

Response 1: Thank you for this feedback. We did test our data for normality using the Shapiro-Wilk method. The results showed that data are not in normal distribution. However, to better understand the meaning of the models when we used the original data, we decided using the original data without transformation.

Comment 2: What the lines inside the study area ranches mean? Are they different "farms"?

Response 2: Thank you for this comment. The lines within the study area are different fields. As grazing and management patterns varied between the fields the internal fencelines were left in.

Comment 3: I don't think Figure 2 is needed, as you comment the results in the text. You can present the graphs in the Appendix.

Response 3: Thank you for this feedback. As suggested, Figure 2 was removed. The Appendix already contained larger figures (Figures A2, A3, and A4) depicting the same results so the text in Section 3.1 was all adjusted to reference those figures instead of the former Figure 2.

Comment 4: I did not understand Table 5, what the data mean. Please clarify.

Response 4: Thank you for this comment. Table 6 (formerly Table 5, Page 6) has been reformatted to communicate the information within more clearly.

Comment 5: Figure 3 presents the results of the best model you showed earlier? The one with red edge and yellow bands? Please clarify that.

Response 5: Thank you for the feedback. The caption for Figure 2 (formerly Figure 3, Page 12, Line 286), was updated to include reference to the specific linear regression model equation that was used.

Comment 6: In line 278 please add r2 instead of r (for consistency with previous results).

Response 6: Thank you for the feedback. Equation 4 on (Page 12, Line 283) was updated to use R²_adj to match the other linear regression equations.

Reviewer 3 Report

Comments and Suggestions for Authors

This study investigates the phenomenon of woody plant encroachment in grasslands using remote sensing technology, a topic of significant ecological importance. The manuscript has a generally sound structure, a relatively scientific methodological design, clear logic, and fluent expression, indicating good overall quality. However, several detailed issues require further modification and improvement. Specific comments are as follows:

Section 2.2: The authors initially generated 1000 random sampling points but ultimately only used 290. Although field surveys are labor-intensive and it is challenging to cover all pre-designed points in practice, the discrepancy between 1000 and 290 is substantial. It is recommended that the authors provide additional explanation regarding the basis and process for sample point selection, for instance, whether points were excluded due to accessibility, site representativeness, or data quality issues. This would enhance the transparency and credibility of the data collection process.

Section 2.2: To further illustrate the distribution of samples across different topographic factor categories (such as slope and aspect), it is suggested that the authors summarize the number of samples for each stratification category in a table. This would help verify whether the stratified sampling effectively covered the heterogeneity of the study area and provide clearer background information for subsequent analyses.

Section 3.3: When establishing the woody cover prediction model, it is recommended to divide the sample data into two parts: a training set for model construction and a validation set for accuracy assessment. This approach would allow for a more objective evaluation of the model's generalization ability and stability, thereby enhancing the rigor and scientific validity of the research conclusions.

Author Response

Comment 1: Section 2.2: The authors initially generated 1000 random sampling points but ultimately only used 290. Although field surveys are labor-intensive and it is challenging to cover all pre-designed points in practice, the discrepancy between 1000 and 290 is substantial. It is recommended that the authors provide additional explanation regarding the basis and process for sample point selection, for instance, whether points were excluded due to accessibility, site representativeness, or data quality issues. This would enhance the transparency and credibility of the data collection process.

Response 1: Thank you for this feedback. In Section 2.2 (Page 4, Lines143-145, 147-149) a more detailed description of the rationale behind the sampling was added.

Comment 2: Section 2.2: To further illustrate the distribution of samples across different topographic factor categories (such as slope and aspect), it is suggested that the authors summarize the number of samples for each stratification category in a table. This would help verify whether the stratified sampling effectively covered the heterogeneity of the study area and provide clearer background information for subsequent analyses.

Response 2: Thank you for this feedback. A table (Table 1) has been added in below Section 2.2 (Page 6) to more clearly illustrate the distribution of samples across the different topographic factors.

Comment 3: Section 3.3: When establishing the woody cover prediction model, it is recommended to divide the sample data into two parts: a training set for model construction and a validation set for accuracy assessment. This approach would allow for a more objective evaluation of the model's generalization ability and stability, thereby enhancing the rigor and scientific validity of the research conclusions.

Response 3:  We agree with the reviewer’s comment. It is common to split data into two groups with one for model development and one for validation. However, we took a different validation approach that is also commonly used – Jackknife validation method.

Reviewer 4 Report

Comments and Suggestions for Authors

This paper used sample data to analyze of spectral characteristics, and used multiple factors to build a model for woody cover. This study is meaningful, but there are some questions:

1. In the 2.2 part, slope was separated into five categories. What are the classification criteria? and following “total 23 unique categories”, there are 4 aspects and 5 slopes, are there 20 unique categories?
2. In the 2.2 part, this study used 10*10m field because of the sentinel-2, but the only 4 bands of sentinel-2 are 10m, is this field enough?
3. In the 2.3 part, “topographic factors including incoming solar radiation”, is this statement accurate?
4. In the 2.4 part, what is the purpose of “using a 3*3 bilinear interpolation”?
5. In the figure 2, it is clearly no correlate in those factors, this figure is redundant.
6. In the table 5, making it difficult to identify the information, what is the meaning of “number-number”?
7. In the 3.3 part, the model has a strong specificity. Maybe it can only use in this area. So you should Validate the applicability of this model in other regions.
8. In the figure 4, there are woody cover. But in the end of the introduction part, the objectives are map woody cover at different progression stages. This figure can not reflect the different progression stages.

Author Response

Comment 1: In the 2.2 part, slope was separated into five categories. What are the classification criteria? and following “total 23 unique categories”, there are 4 aspects and 5 slopes, are there 20 unique categories?

Response 1: Thank you for this feedback. That is correct, the line in section 2.2 (Page 4, Line 140) has been updated to correctly read “20 unique categories”

Comment 2: In the 2.2 part, this study used 10*10m field because of the sentinel-2, but the only 4 bands of sentinel-2 are 10m, is this field enough?

Response 2: Thank you for this feedback. As per the Editor’s comments, we have removed the early reference to Sentinel-2 in paragraph 2 of section 2.2 to reduce confusion. We did not make use of Sentinel-2 imagery in this study.

Comment 3: In the 2.3 part, “topographic factors including incoming solar radiation”, is this statement accurate?

Response 3: Thank you for the feedback. We agree that this was not an accurate statement. The statement on Page 7, Line 169 has been updated to read “Other environmental factors including incoming solar radiation” instead. One other iteration of generalized use of “topographic” was changed in Section 2.6 (Page 9, Line 206).

Comment 4: In the 2.4 part, what is the purpose of “using a 3*3 bilinear interpolation”?

Thank you for the comment. In this case the 3*3 bilinear interpolation helps to generalize the spectral data from the 9x9-meter pixels to better match the 10 m x 10 m sampling units.

Comment 5: In the figure 2, it is clearly no correlate in those factors, this figure is redundant.

Response 5: Thank you for the feedback. The figure has been removed.

Comment 6: In the table 5, making it difficult to identify the information, what is the meaning of “number-number”?

Response 6: Thank you for the feedback. Table 6 (formerly Table 5, Page 6) has been reformatted to communicate the information within more clearly.

Comment 7: In the 3.3 part, the model has a strong specificity. Maybe it can only use in this area. So you should Validate the applicability of this model in other regions.

Response 7: This is a very good comment. For the next step of our study, we will test the applicability of this model in other regions.

Comment 8: In the figure 4, there are woody cover. But in the end of the introduction part, the objectives are map woody cover at different progression stages. This figure can not reflect the different progression stages.

Response 8: Thank you for this feedback. We agree. The language used on the introduction and conclusion have been adjusted to refer to abundance levels instead (Page 3, Line 93)(Page 7, Line 179)

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The author revised the manuscript well.