Age Identification of Farmland Shelterbelt Using Growth Pattern Based on Landsat Time Series Images

Round 1

Reviewer 1 Report

Comments for author File: Comments.pdf

please see the attatch file.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

It was my pleasure to read and review the manuscript entitled “Age identification of farmland shelterbelt using growth pattern based on Landsat time series images”. The authors have used the NDVI-based fractional-cover (acquired from Landsat time series data) of farmland shelterbelt to discriminate growth pattern among farm shelters in Changchun China to extract the age of the shelterbelt. This is a rather simple approach, and as claimed by the authors- universal too, compared to complex regression models which often need modification to apply to new spatial and temporal dimension. Authors work to compile a long time-series between 1984 and 2021. I however, see some major shortcoming in the analysis and ambiguity in write-up that needs to be addressed before the manuscript gets to the publication phase.

Major Comment:

Reporting the increased accuracy with the increase in permissible error appears to be an overwork (increasing accuracy by compromising on permissible error) unless there are clear application areas where having different values of permissible errors can benefit.

Line 485-487: If the phenology of the crop and shelter leads more obvious difference during end of September to the middle of October, why authors chose other months (early May to late June) for this study?

There needs to a significant discussion on figure 9-b. Why the authors think the accuracy, for example of > 30 years age-group is lower than every other age-group until the year 7, and then it starts to overtop?

Discuss the possible effect of Landsat date range (image acquired on Day 126 versus on day > 165) on the FCFS and subsequently on age?

If the canopy width of the farmland shelter belt is 10-30 m, may be the authors could consider how the accuracy increases when they use high-resolution, multispectral Landsat compatible data (such as Sentinel-2 for later years).

Specific Comments

Figure 1: Needs a more descriptive caption. It shows more than just location. What is the second image for? Write that in the caption.

Line 117: Why the image dates are presented in just day of the year format instead of providing what month of the year these days represent and why these particular days were chosen?

Section 2.4.1: Fractional Coverage of Shelterbelt Data:

The authors could have put a sentence to introduce the FCFS first and could have written how the FCFS is related to the index (ExG).

Line 151 and section 3.1: What is dimidiate pixel model?

157: What does ExG index stand for and what it is used for?

Fig 3: Which areas from figure 1 these a and b represent? To ensure the consistency on extent and scale, consider including scale on each figure a and b. Also, consider using location indicator (latitude/longitude) tic or extent indicators on a larger map.

Figure 4 and Figure 5: Write explicitly what percentages of data fall within the Boxes (IQR) of figure 4 and figure 5.

Figure 6: I am not 100% sure if it is the artifact of color or something else, but for me the size of the pixels (level of zoom) don’t seem to be consistent between different years (2013 to 2021). A scale is therefore a needed element in such maps. Also, I see the shelterbelt signal is more prominent in 2018 than in 2019 and in 2021? Why this may be so?

Figure 6-b. IS the curve shown here representative of all the different types of shelterbelts considered in this study? Write that in the section above or below it if it does.

395-396: English requires moderate to extensive editing. An example sentence picked here is “the permissible error increases rapidly in its accuracy from 0 to 3 years” which is confusing and has not delivered the intended meaning based on what is shown in referred figure (figure 9 a-b).

Line 419: How is the main misclassification based on commission error is age 4-15 years while its error is the smallest. Is this a typo?

Line 421: The lowest omission is not 4-15 years based on the data presented on table 1.

Line 436: 4-15 years instead of 4-5 years?

Figure 10: Shelter stands appear outside the borderline of the map? Did author not consider clipping the data and leaving edges inside the borderline?

Figure 11-a: The legend is not correct.

There is inconsistency in citing the references (in some instances I see numbered, and in others I see authors' last name and year).

 

 

 

 

I have made my comments related to English in the general suggestion with some specific examples. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors have put effort to address some of my concerns, but not all and have introduced some new issues as they added new text. 

For instance, one of my major comment was requesting a significant discussion of figure 9.b. Authors described the results but did not discuss thoroughly why  the pattern for example, after permissible age of 7th year >30 year group started overtop everything in terms of accuracy. The likely cause behind such pattern on this figure needs adequate discussion. 

For figure 11, they talk about all the other age group, but has not mentioned about > 30 years age group. 

Writing still needs attention. For instance, one line 85-86 the sentence reads as "For farmland shelterbelts, which are belt-shaped artificial forests, how to obtain ages based on their growth characteristics to be conducted" is not a complete logically structured sentence. Similarly, on line 430, the sentence is written as "The lowest commission of 0.0443 for 4-15 years" which is not a complete sentence either. 

Also, the conclusion needs to point to the future direction the research leads. Author have stated in the manuscript that 16-30 years is one of the two dominant groups and it is the same age group that they stated is most likely to be misclassified. How would they proceed with this if they were to continue the study. 

It has improved, however, still needs some sentences to be rewritten for the sake of their completeness and clarity. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 3

Reviewer 2 Report

OK

Some sentences still need work to render fluidity to the article's readership