A Study of Forest Phenology Prediction Based on GRU Models

Round 1

Reviewer 1 Report

The manuscript presents an approach for forest phenology prediction, from data collection to final prediction. However, it lacks novelty from a theoretical point of view. Only a standard GRU model is used without any modification.

Other concerns are as follows:

1. contributions should be summarized in the last paragraph of Section I.

2. a literature review should be added before "Method" section;

3. more discussion on model designing should be given in Section 2.6;

4. in figure 6 and figure 7, the coordinate labels and legend are too small;

5. in section 3.2, the presented GRU model should be compared with other state-of-the-art methods, e.g. LSTM.

Author Response

The reply is included in the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Authors

Applied Science

Ref: Manuscript ID: applsci-2315852

The work entitled "A Study of Forest Phenology Prediction Based on GRU Models”, by Lichen Zhu, Peng Guan and Yili Zheng

Introduction:

• The aim of this article is to propose a model, called Gated Recurrent Unit (GRU), to predict the Green excess index (GEI) data, and realizes the prediction of forest phenology changes, which can further reveal the future growth of forest phenology

and climate change, providing a theoretical basis for the application of forest phenology prediction.

The authors extensively detail the developed methodology, extensive previous studies are pointed out showing the importance of having increasingly accurate predictive models.

 

 

 

• Notwithstanding the foregoing, I must point out the following:
• 1.- As the lines are not numbered, it is very difficult to point out observations of possible typing errors, spelling signs, etc.

 

• 2.- The abstract needs to be significantly improved. It should be shorter. Parts of the text are unnecessary since they are indicated in the development of the article.

 

• 3.- There is an excessive use of symbols that complicate the reading. That is, for instance: GSI, MODIS, etc, whose meaning is not explained.

 

• 4.- The article shows very clearly the benefits of the proposed model and the statistical control parameters. From which a very good fit is shown. However, a comparative study with other models of objective similarity is not shown.

 

• Suggested improvements:

-       Reduce the length of the abstract.

-       Do not use too many acronyms such as: DIFI, BIFI, SOS, COS, EOS, etc.

-       In Figure 3, change the color of the rectangle inside each image.

-       The caption of Figure 5 does not appear.

-       In Figures 5 and 6 the letters of the axes should be enlarged.

 

 

Author Response

The reply is included in the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

I have no more comments.