Estimating Rangeland Forage Production Using Remote Sensing Data from a Small Unmanned Aerial System (sUAS) and PlanetScope Satellite

Round 1

Reviewer 1 Report

In this revision, the manuscript has been well improved. Thank you very much for the authors considering my suggestions that may not be correct. I have only three suggestions on the revision as follow,

1) Line 527 - 529, the author concluded that on the difference between TOA-NDVI and TOC-NDVI do not affect the data fusion using the STARFM model. Did you evaluate this conclusion by any tests?

2) Compared to the general LUE model, the method used in this study may be site-specific. The reason is that parameterization is highly depended on data collection. How do the authors think about it?

3) Figure S4 may not be able to support the discussion in line 577 - 587.  

Author Response

In this revision, the manuscript has been well improved. Thank you very much for the authors considering my suggestions that may not be correct. I have only three suggestions on the revision as follow,

·       Thanks for your kind words and thoughtful comments. We’ve followed the suggestions and revised the manuscript accordingly.

1) Line 527 - 529, the author concluded that on the difference between TOA-NDVI and TOC-NDVI do not affect the data fusion using the STARFM model. Did you evaluate this conclusion by any tests?

·       Good point.  We rephrased the statement now. We did not directly compare the data fusion results from PS TOA-NDVI and TOC-NDVI. However, the PS TOA-NDVI is mainly used to quantify the temporal change between two dates, and the atmospheric effects would be partially cancelled by differencing if atmospheric condition didn’t change significantly. We do agree that there may be some remaining noise due to lack of atmospheric correction, as acknowledged in Lines 538-540.

“Future release of atmospherically-corrected PS surface reflectance products, i.e., using the concurrent MODIS data, will further reduce uncertainties in sUAS-satellite data fusion methods.”

·       We now rephrased this statement to “In the simplified STARFM, we used the temporal change of PS TOA NDVI, which reduced the noise from TOA reflectance, while the absolute value relies heavily on the sUAS.” (Line 527-528).

·      We now acknowledge that future atmospheric correction for PS reflectance products will potentially increase the accuracy of the data fusion. 

2) Compared to the general LUE model, the method used in this study may be site-specific. The reason is that parameterization is highly depended on data collection. How do the authors think about it?

·       Thanks for the comment. We agree that the coefficients in our LUE scalars may not be applicable directly to other sites. However, we do believe that our model can benefit users investigating similar questions regardless the area. Like other sUAS applications, our model does require collecting sUAS images over the area of interest. With the built image processing and model developing pipeline presented in this paper, one can easily reoptimize the coefficients using their own data.

·       We now added a sentence in discussion to reflect this (Lines 345-346): “The LUE scalars and their corresponding coefficients may be site-specific as they were specifically optimized for our study area.”

3) Figure S4 may not be able to support the discussion in line 577 - 587.  

·       Thank you for pointing that out. We meant Figure S6, the soil moisture time series plots 

·       It is now corrected it in the main text.

Reviewer 2 Report

The authors added some elements to enrich manuscript and I think this paper can now be accepted for publication.

Author Response

The authors added some elements to enrich manuscript and I think this paper can now be accepted for publication.

·       Thank you for taking time to review our work. We really appreciated your comments.

Reviewer 3 Report

This work combines satellite and sUAS remote sensing data to map daily forage production on rangelands. The paper describes in detail the methods used and the results obtained. Organization and writing of the paper are fine.

The main drawback of the paper is the lack of references on related works. There is a good introduction to the existing models on forage production estimation. But there are no references to other UAS-based works, or satellite-UAS data fusion works, which is one of the main contributions. The authors should elaborate more on this.

Other minor comments (mainly format details):

- Figure 1: The map on the right of the image should be enlarged

- Figure 5: For a better comparison, the scale of the y-axis should be the same for both graphs.

- Paragraphs from lines 415 and 547: Correct font and justification

Author Response

This work combines satellite and sUAS remote sensing data to map daily forage production on rangelands. The paper describes in detail the methods used and the results obtained. Organization and writing of the paper are fine.

The main drawback of the paper is the lack of references on related works. There is a good introduction to the existing models on forage production estimation. But there are no references to other UAS-based works, or satellite-UAS data fusion works, which is one of the main contributions. The authors should elaborate more on this.

·       We appreciate your thoughtful comments and cited 6 more references on other sUAS-based production estimation works (Line 113-114):

“A few pilot studies have explored the applications of sUAS technology in monitoring agricultural production for soybean [37], rice and wheat [38,39], barley [40,41], and mango [42].”

·       The implementation of satellite and sUAS data fusion is a piloting topic in the field and reference was rarely available.

Other minor comments (mainly format details):

- Figure 1: The map on the right of the image should be enlarged

·       The map has been enlarged accordingly.

- Figure 5: For a better comparison, the scale of the y-axis should be the same for both graphs.

·       Thanks for pointing that out. We have re-scaled the y-axes to the same extent.

- Paragraphs from lines 415 and 547: Correct font and justification

·       We have edited the font.