Twice Is Nice: The Benefits of Two Ground Measures for Evaluating the Accuracy of Satellite-Based Sustainability Estimates
Round 1
Reviewer 1 Report
Based on the derivation of theoretical formula, this paper puts forward a correction model for evaluating the accuracy of satellite-based estimates. The method proposed in this paper is innovative and interesting. The model is tested by both simulation data and real data.
Some minor problems are as follows:
Lines 59-79 should move from the introduction to the result section.
Line 105：Give the full name of DHS and LSMs, and make a brief introduction of them.
Line 167: “substituting equation (16) into (9) “ should be “substituting equation (13) into (9)”.
Line 246：What is Naïve r2？
Author Response
Based on the derivation of theoretical formula, this paper puts forward a correction model for evaluating the accuracy of satellite-based estimates. The method proposed in this paper is innovative and interesting. The model is tested by both simulation data and real data.
Some minor problems are as follows:
Lines 59-79 should move from the introduction to the result section.
-- good suggestion, we have moved this to the results
Line 105：Give the full name of DHS and LSMs, and make a brief introduction of them.
-- we have added this information
Line 167: “substituting equation (16) into (9) “ should be “substituting equation (13) into (9)”.
-- thanks, we have corrected these numbers
Line 246：What is Naïve r2？
--we have clarified that the naïve r2 is “the direct comparison with ground measures without using equation (14))”
Author Response File: 
Author Response.pdf
Reviewer 2 Report
General comments
It is an interesting work providing a motivating approach to account for noise in ground-based measures when assessing satellite evaluations, trying to improve the assessment of satellite data performance.
However, there are some issues that must be addressed.
Title: I really do not understand the significance of this part of the title. I suggest you delete it; the rest of the title is sufficient and comprehensive; or you should explain it better (you could add "two ground-based measures” or something like this…
Figures and Tables: You should add the figures and tables into the text, according to the guide for authors.
Also, I would suggest you not to write in the first plural.
It is also crucial to explicit what kind of satellite data (and their spatial resolution) were used in these studies that you have collected to test your methodology because I believe that the analysis of the data is very significant for the variation of the results.
It is also interesting to know what kind of VIs were used in these studies. Narrowband greenness VIs? Broadband greenness VIs? The literature has shown that also depending on the crop type they do not demonstrate similar results. Thus, I believe that the clarification of this point is crucial for the evaluation of the noise estimation results.
I think that you should separate Discussion from Conclusions. You must enrich a little bit the literature review of the Discussion (and Introduction I would say) and keep your main findings for the Conclusions.
Specific comments
You can find some more specific comments in the attached file.
Comments for author File: Comments.pdf
Author Response
It is an interesting work providing a motivating approach to account for noise in ground-based measures when assessing satellite evaluations, trying to improve the assessment of satellite data performance.
--Thanks!
However, there are some issues that must be addressed.
Title: I really do not understand the significance of this part of the title. I suggest you delete it; the rest of the title is sufficient and comprehensive; or you should explain it better (you could add "two ground-based measures” or something like this…
--thanks for pointing out. We have changed the title to make the meaning clearer. It is now “Twice is nice: The benefits of two ground measures for evaluating the accuracy of satellite-based sustainability estimates” we feel this is better than the previous more general title of ‘an improved method…”
Figures and Tables: You should add the figures and tables into the text, according to the guide for authors.
--we have moved each figure and table to after the paragraph of its first mention
Also, I would suggest you not to write in the first plural.
-we prefer an active voice to the passive voice, and this is very common in the scientific literature. So we have retained the use of ‘we’ in the article.
It is also crucial to explicit what kind of satellite data (and their spatial resolution) were used in these studies that you have collected to test your methodology because I believe that the analysis of the data is very significant for the variation of the results.
--the yield studies used in this study were all based on 10m resolution Sentinel-2, which we have now specified. The asset and consumption estimates were all based on NightLight data which has 500m resolution, which we also now specify
It is also interesting to know what kind of VIs were used in these studies. Narrowband greenness VIs? Broadband greenness VIs? The literature has shown that also depending on the crop type they do not demonstrate similar results. Thus, I believe that the clarification of this point is crucial for the evaluation of the noise estimation results.
--it isn’t clear which specific studies are being referred to here, but assuming it is the yield studies that utilized Sentinel-2, there were multiple VIs tested. Here we are using the preferred model for each study, which in case of Uganda was based on the MTCI (which uses some of the 20m resolution red edge bands) and in Nepal were based on the GCVI (which uses 10m resolution NIR and green). We have now clarified these differences in the methods section. We also note here (but not in paper) that while the different performance of broad vs narrowband Vis are indeed relevant, it would not affect the general point of the current paper that failing to account for noise in ground data would underestimate the VI performance.
I think that you should separate Discussion from Conclusions. You must enrich a little bit the literature review of the Discussion (and Introduction I would say) and keep your main findings for the Conclusions.
--we have now added some discussion and split conclusions into a separate section
Specific comments
You can find some more specific comments in the attached file.
--we have edited the manuscript according to these comments, many of which were addressed above.
Author Response File: Author Response.pdf
Reviewer 3 Report
Interesting and well written manuscript on how to deal with noisy field measurements to assess the accuracy of remote sensing based estimates. Useful for many to consider. In my opinion the methods section should be improved and clarified, and in addition there are a few things in section 3 that could be discussed further. I have annotated the manuscript with my comments, mainly in sections 2 and 3. I recommend minor revisions.
Comments for author File: Comments.pdf
Author Response
Interesting and well written manuscript on how to deal with noisy field measurements to assess the accuracy of remote sensing based estimates. Useful for many to consider.
--Thanks!
In my opinion the methods section should be improved and clarified, and in addition there are a few things in section 3 that could be discussed further. I have annotated the manuscript with my comments, mainly in sections 2 and 3. I recommend minor revisions.
--we have edited the manuscript according to these annotations, which were all quite helpful. Thanks for the detailed review
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors have addressed all the comments, and the paper is now suitable for being published.
