Optimizing Near Real-Time Detection of Deforestation on Tropical Rainforests Using Sentinel-1 Data

Round 1

Reviewer 1 Report

This is a very competent manuscript.  It is well written and presented.  The tables and figures are informative and professional.  There is a very extensive, almost two extensive, set of appropriate references.  The topics of tropical deforestation and SAR are important and will be of interest to the scientific community.  

The authors are to be congratulated on an extensive amount of data collection and analysis.  The manuscript is however very long and complex reducing to some degree its impact.  There is an excellent literature review.

As in most lengthy manuscripts, there are editorial suggestions for considerable by the authors, several of which follow:

1. It varies by journal and authors but use of personal pronouns (we, our, us) is not typical in most scientific text.
2. Line 18, perhaps not intent as it was completed.
3. There appears to inconsistent use of serial commas.
4. Not all acronyms appear to be defined and then employed.
5. The term state-of-the-art is detracting.
6. Line 98, stabilization.
7. Line 108. Data are plural.  Resolution needs a modifier such as spatial or temporal.  Fine is preferred to high resolution.
8. Line 154, defying? Challenging, difficult?
9. Line 199, works? Perhaps efforts, studies.
10. Line 239, totalizing?
11. Line 240, yr and per year?
12. Units do not appear consistent, meters, ha?
13. Contractions (don’t) are not typical in scientific text.
14. Is Table 5 cited in text?
15. Appendices are not typical. Are they necessary?
16. There are some inconsistent formats in the references. Several article titles are in upper case.

In summary, this is a very competent but lengthy manuscript certainly suitable for Remote Sensing.  The provided comments are all minor editorial issues.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

I would like to congratulate the authors for doing such a nice job doing this research in the first place and for the writing part. I found the manuscript very informative, but lengthy. I think it's not an issue with this journal (no word limitations). 

I found the breakdown of the sections very helpful. The detailed review of the state of the art of technology is really helpful. The materials and methods section was detailed enough to provide the required information appropriately. 

The authors had used plenty of figures in the results section to provide a very detailed and informative presentation of the results. Since there's no limitation for the number of figures, I didn't find it problematic. It provides the full picture. 

Nice job with the discussion and conclusion sections!. 

Please find my comments/suggestions in the attached pdf (commented within the pdf). 

Thank you!

Comments for author File: Comments.pdf

Author Response

Thanks for the kind remarks and for the valuable comments. We have applied all the suggested corrections. 

Reviewer 3 Report

The paper is discussing the impact of spatio-temporal filtering on deforestation detection for the real time deforestation monitoring. I think authors have good knowledge in the topic and the methods applied in the manuscript seems relevant and sound.

Just one comment on the methods is that, since their classification methods are heavily dependent on the thresholds from the statistics from samples, it is important to make sure that the numbers of samples are large enough to provide a stabilized threshold. Readers may be curious to know how many samples are ‘enough’. A plot could be also helpful to show with which number of samples, the threshold stabilizes etc. 

Otherwise, my comments are mostly about the way the manuscript is written.

Too much details about the methods:

3.1.2 sampling spaces

Authors may provide a shorter description on sampling and provide most of it as appendix.

4.1, 4.2 maybe authors can you just summarize this section in one sentence and move this section to appendix.

There are some parts the manuscript not clear:

350 “In this study, we applied the Spatio-temporal filtering procedure proposed in…” -> I know that this is for removing noises but the purpose of doing this should be presented first.

Figure 8 two modeled component -> forest and non-forest component?

Figure 9, why not include original time series in this analysis? Readers would be curious to know if filtering etc improved the separability from the original time series.

Figure 11. Not sure what information I should get from this figure. And again, why not include original time series?

Table4,5,6. no results from original time series? Wasn’t the purposes to show how much the filtering and processing improve the classification?

Some typos and miss spells:

144 changes ->changes

177 what it means that “C-band will degrade quicker, being unusable after 5-10 days”?

209 “there is an” -> “there is no”?

248-249 what do you mean by this, did you take a more complex procedure?

501 two year are collections - > two year collections are

Author Response

Thank you for your comments. Please see the attachment for the answers.

Author Response File: Author Response.docx

Reviewer 4 Report

This manuscript examines the application of SAR data for DETER, a long-time operational Near Real-Time (NRT) deforestation detection system. The SAR-based NRT system is particularly valuable in areas that are frequently covered by clouds. Therefore, this study will provide useful information to many researchers. Therefore, I believe the manuscript is worth publishing. However, there are some confusing parts, and I recommend the following minor revisions.

1. [line 13] Subscript the 2 in "CO2".

2. [lines 54-62] This description does not fully demonstrate the novelty of this study. It would relate to the history of previous studies described in Chapter 2, and the authors should describe which methods this study employed and what novelty this study has among such studies (perhaps at the end of Chapter 1 or Chapter 2).

3. [Table 1] There is a mix of items that are evaluated with + and -, and items that have specific numbers. I think it would be better for all items to be evaluated with + and -, and for some items to be noted with specific values.

4. [line 131] PALSAR-2 is onboard "ALOS-2".

5. [line 132] PALSAR-2 usually observes HV polarization, not VH.

6. [line 224] This table does not have number and caption, so describe them.

7. [line 240] "yr-1" and "per year" are overlapping words.

8. [line 354] Perhaps "lp(x,y)" is a miswriting of "lk(x,y)".

9. [line 360] "ththe" is a typo.

10. [line 395] "the a" is also a typo.

11. [section 3.2.4] I do not understand what is the truth in the validation described in this section. Please describe how the authors have validated it.

12. [line 557-559] I'm not sure what the authors would mean by Figure 9: do they show that 1-yr and 2-yr are almost the same?

13. [line 650] Isn't "MLC" a miswriting of "ALC"?

14. [line 654] At the end of this section, I think the authors should conclude whether the MLC or ALC method is superior.

15. [line 670] The date of the start of DETERSAR operation is described, but also describe the year.

16. [line 869] The paper by Koyama et al. was published in 2017, not 1997.

Author Response

Thank you for your comment. Please see the attachment for the answers.

Author Response File: Author Response.docx