Uncertainty Analysis of Remote Sensing Pretreatment for Biomass Estimation on Landsat OLI and Landsat ETM+
Round 1
Reviewer 1 Report
This is an interesting treatment of an interesting issue. However, it also has a lot of problems and needs improvement in a number of areas (the Results section, in particular)
I’d prefer a more specific Abstract – its very general; “the effects of three pretreatments on uncertainty of biomass estimation were objectively quantified” – give details.
Line 121 “The difference in RE between each stage and the previous pretreatment stage is the uncertainty of biomass estimation.” Should be in the Materials and methods section, not Introduction.
Section 2.1 "Rolling mountains" with narrow steep slopes?
Section 2.3 I’d be interested in a rough idea of what the NFI biomass values are – ie mean value and range.
Section 2.4 (Random Forest modelling) copies the similar section (2.6) in “Mapping Tree Canopy Cover and Aboveground Biomass in Sudano-Sahelian Woodlands Using Landsat 8 and Random Forest” (Remote Sensing, 2015) almost word-for-word and needs to be redone!
Results: explanation needs to be improved and parts rewritten. Sentences such as “the low-value frequency reduced” (Line 260) is not sufficiently clear.
What was the range for observed and predicted biomass?
You say: “(b) The abnormal values decreased.” (line 271) How do you know this: its not immediately apparent to me from comparing Figs 3b and 7b. For example, there are more spikes of >120 in Fig 7 than Fig3 (Landsat 7). I think you have to justify this argument. I’m not sure these R^2 diagrams are a clear way of showing this and you should consider alternatives.
“although the OLI precision was slightly better owing to a larger data bandwidth” Doesn’t seem like a big difference to me. (Line 211)
You spend more time discussing training sample results than test sample results, both in Results, Discussion and Conclusion. It should be the other way round.
Figures: Fig 1 : inset map showing location of study area would be useful Do you need to show training sample R2/MAE (ie (a) in Figs2-9) in a graph? You could give all the values in a table, and you could merge the R2/MAE figures(ie Fig 2 and Fig3), and halve the number of figures. I think it would be helpful for some readers if you were to give a brief explanation of the diagrams at the start of the Results section.
The standard of English is generally good, but there a few minor corrections.
Line 41 insert as “especially ^as^ it plays”
Line 45 Delete t “of t carbon sequestration”
Line 59 Do you mean “poses a spatial problem”
Line 75 Delete comma “Previous, research”
Line 100 do you mean “pretreatment is limited”
Line 134 change to “obviously cut”
Line 135 “narrow slopes are steep and the ridges clear”
Line 135 “ridge is clear “ do you mean clear of trees – open area?
Line 138 “delete “for” “cloud free Landsat”
Line 148 delete “airborne”
Line 148 damaged by Scan Lines Corrector (SLC) failure”
Line 199 "calculating the difference between the relative error of the previous stage pretreatment results, summary results, comparison and analysis." What does the last 6 words mean?
Section 2.3 Info on the SLC failure should come first tin this section – reorder sentences.
Line 257 “ranges?
Line 320 delete a “reporting a about 3%”
Line 357 “helpful ^in^ reducing the uncertainty”
Line 359 Remove comma “Although, we predicted”
Line 360 “In this study, only one correction model chose of each pretreatment.” Don’t know what this means
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 2 Report
The Manuscript discussed on improving the accurate quantification of biomass using different methods and statistical models. The manuscript was written well and presented the data clearly. I recommend the article to publish after a minor revision.
1. Figure 1. The numbers on the axis are not clearly visible. Also, the legend (text) in the figure is not clearly visible. Please increase the font size to make it clearly visible. Similarly, increase the axis font size and legend text size for other figures 2-9.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Although the article deals with the current topic of biomass estimation from satellite data, it does not bring much new knowledge. Although I am not an expert in English grammar, at first glance it needs to be corrected by a native speaker, since the article contains many typos and errors, especially in the order of words. Also some informations from chapter Introduction should belong more to methodology (lines 115 - 124).
Line 45: the carbon sequestration
Lines 138-142: Sentences are not clear for reader and should be rewritten.
I see a lack of information in methodology, e.g. description of NFI in China and on the other hand, some are redundant (description of Random Forest technique).
Line 147: We assumed that the data from sample plots was the true value of biomass..... Somewhat bold statement, because any parameters of NFI are not mentioned, especially precision of sample plot location and measured parameters.
It is not also clear what kind of spectral indices were finally used for AGB estimation.
The article can be published after complete revision. For example, it is not clear why the authors used 2 datasets when ETM + on Landsat 7 is corrupted. As a whole, I recommend to completely rewrite, supplement the methodology and especially check with a native speaker.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Improved, however, still a few corrections required (see below)
Line 30-36 To avoid confusion I'd give only the uncertainties for test or training sample, rather than giving both. If you do give both, you must explain what's going on.
Line 34 "atmospheric correction"
Line 162 "2.4 Methods" not very informative heading. "Pre-processing stages?
Lines 170-172 These last 2 lines would be better incorporated into 2.8 (Variance) section
Line 164 "based on the two indices described above." Not obvious to me what you are referring to here
Line 171: should be R^2
Line 183: Change to "All the vegetation indexes"
Line 185: "NIR is NIR band" is not informative. Say Near Infra Red.
Line 218: Always give abbreviations on first appearance SD(standard deviation)
Lines 224-226: 15 of what? (Rounds?)
Lines 226-227 "Compared with Figure 2 (b), we found that the MAE corresponding to the low R2 value of the test sample and this was abnormal." Still not clear what this means
Line 267 "number of rounds that values" should be "number of rounds with values"
Line 284 "^For^ both OLI and ETM+, the mean value was"
Line 290-291 "after atmospheric correction using" should be "after using atmospheric correction"
Line 294 "31.95% with a 4.41% lower" should be "31.95%, about 4.41% lower"
Line 296 increased significantly
Line 297 "rounds that values" should be "rounds with values"
Line 309 as for Line 297
Line 309 "that signify an increasing trend of accuracy" should be "signifying an increasing trend of accuracy"
Line 317 "than ^for^ atmospheric correction"
Line 317 "MAE range reduced to 20.47~20.01 t/ha with a mean value of 20.74 t/ha" Is the 20.01 value correct?
Line 320 "with" for "that" as for Line 297,309,etc
Line 394 "more attempts in ^the^ future to understand"
Line 395 "different correction model^s^"
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Authors improved the article in the terms of regarding changes but still in some parts it could be better descibed e.g. for me it is still not understandable enough how authors combined vegetation indices for AGB calculation. However the article brings new approaches for AGB estimation and I apprecciate especially differences in results with different corrections and data processing and accept it after minor revision. It would be nice to describe at least the tree species in general and explain how were vegetation indices combined for AGB estimation.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.