Comparing the Utility of Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN) on Sentinel-2 MSI to Estimate Dry Season Aboveground Grass Biomass

Round 1

Reviewer 1 Report

This paper compares the performance of ANN and CNN in estimating aboveground grass biomass in dry season land, and concludes that CNN is superior to ANN.

1) It is not clearly stated in the text whether all the features listed in Table 2 were used in the modeling process. Suggest conducting a correlation analysis between characteristics and biomass.

2) Line 280, I am sure it should be Figure.5. Please correct it.

The quality of English writing should be improved.

Author Response

1. All vegetation indices were inputted into the neural network models. However, only significant indices based on average impact was used for model development. (highlighted in yellow).

2. Corrected (highlighted in yellow)

Reviewer 2 Report

Please see the attachment.

Comments for author File: Comments.pdf

Author Response

1) There are no known similar studies at a local or regional scale (ie. Comparing the performance of neural networks to assess grass biomass). Relevant studies have been provided in detail in the discussions section. In addition, a paragraph linking biomass estimation to climate change was added (highlighted in green)

2) Added (changes highlighted in green)

3) The purpose of Figure 2 is to provide the reader with a visual aid to gain a better perspective and context of the in-field conditions. Hence, I would motivate that it is useful to show in the methods section.

4) Added paragraph on the implications of applied technique (highlighted in green).

Reviewer 3 Report

The article is devoted to current topics. I recommend it for publication after minor blemishes have been corrected.

A link to the sites line 172 and line 185 (Table 1), line 209 should be given according to the rules of the journal in the form of a number and mentioned in the literature.

Figure 2 should write the general title of the figure and the designations of the parts of the figure should be given after writing the name.

In Figure 7a, the legend overlaps part of the vertical axis.

Author Response

1) Links provided in the form of a number and referenced in the references section (changes highlighted in blue)

2) Amended (changes highlighted in blue)

3) Corrected (changes highlighted in blue)

Reviewer 4 Report

Dear Authors,

In the paper Comparing the utility of Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN) on Sentinel 2 MSI to estimate dry season aboveground grass biomass using vegetation indices, CNN and ANN methods were compared. The research was carried out based on an image from the Sentinel 2 satellite. The description shows that they used data from one image for the analyses. The reader may be unsatisfied, repeating the analyzes for different or several terms could be appropriate in the context of controlling the analyses. However, taking into account the main goal of the work - and the search for the answer which neural network will predict biomass better - the material is sufficient. Sections 2.5.1, 2.5.2 and 3.2 generally present the theoretical foundations of ANNs and CNNs. In the content of the work, apart from the general table 3 with parameters for two models, we will not find any details. Among others information about the trained data set. How the input data was prepared and the training parameters defined. For example. whether additional transformations, data transformation, etc. were used. In the case of CNN, the method was a simple matrix transformation of the image and extraction of features. In my opinion, the information should be detailed and supplemented.

Author Response

Additional information was added. This is highlighted in pink. 

Round 2

Reviewer 2 Report

The manuscript is sufficiently improved.

Best wishes,