Impact of Modeling Abstractions When Estimating Leaf Mass per Area and Equivalent Water Thickness over Sparse Forests Using a Hybrid Method

Round 1

Reviewer 1 Report

Overall, this is a fine, interesting paper. My main complaint is that the Materials and Methods section, while detailed, is not always clear, and that values suddenly appear, without any explanation of why they have been chosen, or where they come from.

 My suggested corrections are divided into two, first more major points, then minor corrections, typos, formatting…

Major Points:

Abstract/Discussion – lns 11-15. Ln 320-334 Ln432 Given RMSE values are surely x10^-4, or in different units to those given

Section 2.3 & 4:  AVIRIS-Next Generation hyperspectral data – would be useful to give the spatial resolution here, especially as later in the Discussion you discuss remote sensing resolutions – how do these compare to your current study?

 

Ln 197 Where does the 60% empty voxels value come from?

 

Ln203 “Other overstory characteristics corresponded to those…” What characteristics – be more specific.

 

Table4 canopy height values (9.3 and 13.6) Where do these values come from? Why have you changed canopy height from SFR to Variation

 

Table5: Presumably middle column values ie for LAI,ALA EWT, N apply to all columns:is this right?

 

Table 5: I’d like more clarity on why these values were chosen i.e. EWT and LMA max value of 0.025. It’s different to those of the (admittedly few) field study values. Where does structural parameter value N come from? Why was this range of LAI (0.3-4) chosen?

 

Table 5 You use the crown height/tree height values from Table 1 for TZ but not for SJER (where does 5.7m crown height come from?)

 

The “sensitivity/synthetic” column in Table 5 is not very helpful – so many “database dependent”, which is not very informative. Does this mean SFR/DETAIL databases? If so, why not give values in 2 columns for SFR and DETAIL?

 

Materials and Methods I think clarity is needed – as the different number of models you mention– SFR, DETAIL, TZ, SJER, synthetic,sensitivity, variation – I’m having trouble keeping straight what is used for what. Perhaps a summary section at the start might help, where you clearly state in 1 section the names and number of all the different databases used.

Table 5 Units for cell dimensions and scene dimensions – 0.4m2 x 0.4m2 ?

 

Section 2.8, Ln 232 “three different training databases”. Then you detail 4 databases (SFR,DETAIL,SJER, TZ)

 

Ln 248 “into account the results presented section 3.2” What results ? and how where they used? You give no details. If you are using results from later in the paper to inform, you have to at least briefly summarize what they were.

 

Ln257-259 “For the SJER database, crown diameter, tree height and crown height were scaled so that crown cover in the DART scene remained consistent, keeping the proportions given in Table1.” I’m afraid I don’t follow this line at all. Why just SJER and not TZ? This is the first time that you’ve mentioned “crown cover”. Is it the same as “canopy cover”? “proportions given in Table1” – Table 1 doesn’t include proportions. What exactly does “consistent” mean? Please make this line clearer.

 

Results/Discussion – there is only minimal/no mention in the text of either section of the field study results. I think it would be good to say something as they are extensively mentioned in the Material and Methods section. The most detailed mention in the text is at the end of Section 2 (lines 280-281)

 

Appendix Figure A1 I don’t understand this figure. I think caption may be faulty, with (a,b,c,d)  and (e,f,g,h) repeated description in the caption. Why are (a) and (e), (b) and (f)…. repeats with regard to axes, but show different results. I think you need to explain this diagram further.

 

 

Minor corrections/suggestions given below

 

Ln 12 However, EWT estimations were not affected by…

Ln14 obtained accuracies (RMSE of 31…)

Ln32 fire risk is likely

Ln70 knowledge

Ln80 providing acceptable results

Ln86-87 confusing line do you mean “effect tree modelling abstractions ^have^ within a RTM on crown reflectance, to identify spectral…”

Table 2 – I can only see 11 markers in Fig1 – 5 in TZ and 6 in SJER. You say 12 samples in this Table – I also note there are also only 11 data points in Fig 7b) and d)

Figure 1 On the right hand image, it would be good to have San Joachin marker in blue, to help reader see quickly which is which

Ln191-192 “up to a thousand trees” prefer more specific numbers

Ln248 “results presented ^in^ section 3.2”

Table 5 caption variable parameters

Ln287 I’m not sure “Similarly” is the right word to use here, given that the effect is the opposite to that previously discussed…

Ln291 variation

Ln330 “equal to the DETAIL scenario”

Ln333 delete “comprised”

Fig7 Why are the markers in (b) and (d) red and blue? Does this refer to the 2 different study sites – if so, you should say so, and say which is which

Ln368 presented ^in^ Table 7

Ln375 “if crowns are not modelled”

Ln382 databases

Ln387 trait

Ln413 give all abbreviations (signal to noise ratio?)

Ln 425 “allowed the rejection of the most biased bands

Ln426 However, EWT estimation …

Ln429 “sensible” Do you mean sensitive?

Ln439 “suggested rejecting the…”

Ln458 “could allow the effective isolation of tree crowns”

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

It is interesting to investigate the relation of spectral bands and physiological parameter combined with 3D structure construction and machine learning based data derivation. In general, I think the research is comprehensive and significative.

Please see my detailed comments below:

The typeface in different figures could be unified, especially for figure 5\6\7

The full stop could be deleted in the title

Line 2: ‘vegetation process’ the expression could be improved

Keywords: ‘EWT and LMA’ are not included.

Line 23-24: ‘including as they do,’

Line 42: …physically-based methods and hybrid methods)

Figure1, those three maps could contain the basic elements like proportional scale, longitude and latitude(with the unit of °), compass and so on.

Table 1: please check the number for SJER Crown diameter, 16?

Line 126: the reference could be added for the EWT and LMA computational formulas even well-known.

Line 184:’the DART pixel size was 40 cm’, why?

The range of EWT in figure 8 is different.

Line 368: in Table 7

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have responded in detail to all my comments, and I have no further issues. Best of luck!