Exploring the Applicability of Regression Models and Artificial Neural Networks for Calculating Reference Evapotranspiration in Arid Regions
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe manuscript is well structured, admirably clear, and succinct, and it was a pleasure to read. Figures are well presented. However, the literature reviews and new papers (since 2020), are sparse (especially in the introduction). I, therefore, recommend the paper be revised before it can be published (see my comments per section).
Abstract:
- It should be mention that the reference ET0 in this study was calculated based on FPM method.
- The abstract introduces the two modeling techniques (stepwise regression and artificial neural networks) and briefly mentions that they were applied to climate data from Egypt. It might be helpful to include a sentence describing the length of the dataset, as well as numbers of stations.
- The abstract effectively highlights the key findings, specifically that the artificial neural network (ANN) model outperformed stepwise regression, as evidenced by the high coefficient of determination (R²) and low mean absolute percentage error (MAPE). However, it would be beneficial to provide specific values for R² and MAPE in the abstract itself, as this would immediately convey the level of accuracy achieved by the ANN model.
Introduction
- I highly encourage authors to use this following citations (be sure, I am not one of the authors of this papers):
** https://doi.org/10.1134/S2079096120040150 (Evaluation of Estimation Methods for Monthly Reference Evapotranspiration in Arid Climates; Highly relevant to your study).
** https://doi.org/10.1016/j.ejrh.2022.101259 (Reference evapotranspiration estimation in hyper-arid regions via D-vine copula based-quantile regression and comparison with empirical approaches and machine learning models)
** https://doi.org/10.1016/j.jhydrol.2023.130021 (A multi-decadal national scale assessment of reference evapotranspiration methods in continental and temperate climate zones of South Korea; This new paper is a national assessment of alternative models for estimating ET0)
** https://doi.org/10.1002/joc.7894 (High-resolution reference evapotranspiration for arid Egypt: Comparative analysis and evaluation of empirical and artificial intelligence models; Its in Egypt and authors did not include it!)
- L 37-39: Add a reference to support this.
- L 49-51: Add a reference to support this.
- L 97-107: Summarize it to max 4 lines
- L 107-119: Summarize it to max 4 lines
- L 119-130: Summarize it to max 4 lines.
- L 87-130: Its a literature reviews, so at the end of these reviewers, authors should highlight the conclusion of this review and the research gap(s).
Materials and Methods:
- Fig.1: Improve the quality of this figure.
- L 148: Which version?
Results:
No comment.
Discussion:
- Overall, its a short discussion. Add more supporting statement there. I am familiar with some papers, and authors should enrich this part.
- At the end of discussion, I would like to see research limitation and future directions.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsDear Sirs,
The article you presented is interesting, it takes up the needed issues to describe- Investigation of the usefulness of regression models and artificial neural networks for calculating reference evapotranspiration in the dry region.
Neural modeling is gaining importance in the development of agricultural science, and the stepwise regression method is a good tool for comparison. I have a few comments on the neural network methodology and discussion of the results. Please see the following questions and answer.
1) In the Abstract section, I suggest adding a brief information on how many years of research were included in the development of the models.
2) In the Introduciotn section, please add one paragraph in which you characterize the methods of stepwise multiple regression and ANN.
3) In the methods section, please characterize in detail the study site-microclimate, soils, etc.
4) Please specifically list what meteorological data from weather stations you used for your research (you can add a separate subsection in Chapter 2).
5) The methodological part must include information such as the method of learning neural network, what type was tested, how many variables were used, what was the accepted division of sets: learning, validation, test, what part of the data was used to build, and what part was used to validate the model.
6. Please specifically state in the results section what type of neural network was built, how many neurons were in layers and why, etc.
7. What value of MAPE error is acceptable to you.
8. Plase refine your discussion of the results - include at least 10 publications from the last 10 years on similar topics.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsOpinion:
The research study focused on developing a model to estimate reference evapotranspiration (ET0) using different climatic variables as predictors. ET0 is a crucial parameter in agriculture and irrigation management, especially in arid and semi-arid regions. The study evaluates two modeling techniques, stepwise regression, and artificial neural networks (ANN), to determine which one is most effective for calculating ET0. The research uses climate data from Egypt's entire climatic station for model development and testing. The subject covered in the manuscript is suitable for the journal. The abstract is well written. The introduction section is processed at sufficient level. The materials and methods section is adequate. The MS is properly organized, clear, and concise. The article's organization and language are generally good. The article effectively discussed the research findings. Moreover, conclusions are supported by the results. The manuscript deserves to be published after minor revision.
Minor revision:
l The Abstract needs to highlight what is new and achievements.
l I invite them to highlight the novelty statement in their manuscript to show to readers that the current work brings a novel idea to the existing knowledge.
l In line 397 absolute percentage error (MAPE) of 2.7.
l Please add your future recommendations.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsI accept.