Estimating the Routing Parameter of the Xin’anjiang Hydrological Model Based on Remote Sensing Data and Machine Learning

Round 1

Reviewer 1 Report

remotesensing-1874510

Title: Estimating the Routing Parameter of the Xin’anjiang…

1.      The subject is interesting, but some major revisions should be considered. Also, please, start revision in MS-Word> Review menu, while the track change is on for the rapid referee in the next stage.

2.      The quality of Figure 1 is low. The numbers are not seen correctly.

3.      In Figure 2, some abbreviations are not mentioned in its title. Titles should be informative and complete.

4.      In Table 2, add a column that mentions sources for each row.

5. In Figure 5, the title should be more precise. What do the numbers in each house represent?

6.      Figure 10 of the maps will be more understandable by adding the waterway network and Hillshade. Some graphs are not good and clear too. Cite and refer to the following articles as a template: “Assessing the expansion of saline lands through vegetation and wetland loss using remote sensing and GIS” for some ideas and indexes in land changes.

7.      Some recent photos from the region can promote clarity. For this refer to this: “Ecotourism and socioeconomic strategies for Khansar River watershed of Iran”

8.      The abstract should briefly state the purpose of the research, the principal results, and major conclusions. An abstract is often presented separately from the article, so it must be able to stand alone. Add some stronger conclusions in it.

9.      It is suggested to present the article's structure at the end of the introduction. At the end of the introduction add a para including 1-Gaps in the backgrounds you try to fill them, 2-your novelty and unique aspects 3-Hypothesis 4-Objectives.

10.   Please make sure your conclusions section underscores the scientific value added to your paper, and/or the applicability of your findings/results, as indicated previously. Please revise your conclusion part into more detail. Basically, you should enhance your contributions, hypothesis retain/reject, limitations, implications/applications, advantages/disadvantages, policies, underscore the scientific value added to your paper, and/or the applicability of your findings/results and future study in this session.

Author Response

Please find attached our responses to your comments.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript reports the use of classical machine learning algorithm (gradient boosting tree) to predict the hydrological model parameters. It is an interesting read and is in the right direction towards use of data-driven approaches within the water domain. I have following comments for authors to improve the quality:

1.     A detailed literature review section is missing. Authors need to identify the studies from literature where machine learning regression approaches have been used to address the water-related problems. I can understand that there might not be any literature where hydrological parameters are predicted using ML, however, from applied perspective, there are many where similar problems have been addressed i.e., given a set of input features, predict a target variable. Authors should review such studies and critically analyse in a separate section. Some candidates for such review could be but not limited to:

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018WR023325

https://www.sciencedirect.com/science/article/pii/S0022169405003641

https://www.sciencedirect.com/science/article/pii/S0022169420302407

https://www.tandfonline.com/doi/full/10.1080/1573062X.2022.2075770

2.     A fundamental question about the selection of specifically the boosting tree models and why not any other model? There are deep learning models also available which are considered more robust than conventional models, why those ANN, 1D-CNN models were not implemented or compared. Also, it would have been much better towards enhancing the contribution of research that authors would have implemented at least common models from both conventional and deep learning domain and compared those to highlight the best one with proper justification. As of now, it seems like a random choice and gives an impression that most important part of research (model selection) has been ignored. 

3.     I would suggest authors to include a correlation heatmap between the selected input features and target variable to demonstrate the relevance of selected features. Refer to one of the papers suggested in point 1 to understand the relevance of heatmap and its explanation.

4.     Authors are suggested to clearly tabulate the models 1-5 in terms of what parameters each model is developed based on. As of now, it is also unclear that why 5 models were selected? What was the intuition behind developing exactly five models?

5.     Figure 6, not clear, way too cluttered. I would suggest authors to add scatter plot for each model separately and mention the R^2 score on each scatter plot to better understand the performance. 

6.     Figure 9, does not convey any information that can be interpreted. It is suggested to redraw or explain it in a better way. Again use one scatter plot for each model. Do not clutter only one graph.

7.     Figure 11 and Figure 12, not clear at all. Could have been presented in a much better way. First, may be log plot may be useful to capture the y-axis variations. And again use one plot for each of three variants. 

8.     Authors need to address the implications of the research. The discussion should show how your findings fit with existing knowledge, what new insights they contribute, and what consequences they have for theory or practice. Do your results agree with previous research? Are your findings very different from other studies? Do the results support or challenge existing theories? Are there any practical implications? Your overall aim is to show the reader exactly what your research has contributed and why they should care.

        9. It is suggested to add the limitations of your research and propose the potential future directions of the research based on the findings. 

Author Response

Please find attached our responses to your comments.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Authors have addressed most of comments made in round 1 to my satisfaction.

Author Response

We would like to thank you again for positive feedback.