Spatial and Temporal Biomass and Growth for Grain Crops Using NDVI Time Series

Round 1

Reviewer 1 Report

Good for the publication in the present form

Author Response

Please find attached pdf.

Author Response File: Author Response.pdf

Reviewer 2 Report

In this study the authors present their approach in handling/creating long time series of NDVI measurements derived from, mainly, S2 images in order to estimate above ground biomass in croplands. The research design is based on extracting important phenology information for crop growth in parts of Australia and could be of interest for the readers and potential end users. Another important aspect in their methodology is the inclusion of additional sources of data to complement or fill gaps in the time series. The authors make use of various field and satellite datasets along with established methods to extract the required metrics.

While the research design follows the initial objectives defined in the introduction, my main problem was the overall structure of the paper and presentation of methods and results, which led me several times to go back and forth in pages or to the appendix in order to comprehend the implemented steps. Consequently, i had to wonder what was the focus of this study - was it the extraction of regression metrics, the gap filling of the spectral matching of S2 and Planet images.

I would suggest Appendix B to be included in the main methodology text - i do not see why it should be in appendix. Alternatively, figures 6 or 7 could be placed in the appendix as supplementary material.

Another important flaw is not including the study area(s) in a figure and a brief description about these lands. Following this, i would expect more details on the amount of data used (number of images, dates, etc.). Who provided the field data? A separate Conclusions section is also missing!

The training/calibration and validation dataset terms and usage appears confusing. Some tables are in main text while others are in appendix and this does not help. Here, an addition of a general methodology flowchart would be of great help for the reader to understand the training (calibration/validation) and the separate validation exercises followed.

Why is the log10 transformation applied? Please explain.

Lines 227-229: Why are the 31 and 76 day thresholds are selected? Please clarify.

 

 

 

Author Response

Please find attached pdf.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript titled ‘Spatial and Temporal Biomass and Growth for Grains Crops 2 Using NDVI Time Series’ demonstrates a framework for estimation of AGB and crop growth rates under different soil background and crop phenology based on Sentinel-2 imagery using adjusted summed NDVI, and analyzes the impacts of time series gaps of the imagery on the estimation results. The topic is interesting and the methodology is simple but basically available.

1.     Introduction: the background is presented but the scientific issues are not identified clearly. What is the novelty of this research?

2.     It’s better to insert the appendix A and B into the corresponding parts of the text.

3.     Materials and Methods: some figures should be added to demonstrate the different remotely sensed imagery and different crops distributions. The acquisition time and orbits of the satellite imagery used should be given.

4.     Methods of calculating sNDVI: The greenness threshold NDVIgt is directly related to a, which was set to 0.10, why? This threshold is a key parameter for AGB estimation.

5.     Some necessary notes for the parameters should be added in table1.

6.     According to the appendix A, the crop species involved in the training and validation sets are slightly different, but in table1, they seems to be the same, and some species without shown in appendix A have the validation results, which is also inconsistent with figure A1.

7.     More AGB estimation maps of different sites and crops should be given in the results.

Author Response

Please find attached pdf.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have performed accordingly the required modifications and additions in their manuscript. I would still insist on a figure with the study area(s). Even at coarse scale, there are ways to present the test sites. In remote sensing there are numerous global studies that contain this information...

Author Response

Please see attached.

Author Response File: Author Response.pdf

Reviewer 3 Report

All my comments were addressed, but I have one more suggestion for the conclusion. 

The accuracy of the models should be given in the conclusion, and the statements  in Line 489-496 could be moved to discussion section. 

Author Response

Please see attached.

Author Response File: Author Response.pdf