Assessment of Machine Learning Models for Remote Sensing of Water Quality in Lakes Cajititlán and Zapotlán, Jalisco—Mexico

Round 1

Reviewer 1 Report

The manuscript titled Assessment of Machine Learning Models for Remote Sensing of Water Quality in Lakes Cajititlán and Zapotlán, Jalisco - Mexico, presents interesting topics about water quality and analyzed the satellite data and using ML for prediction. Although manuscript is well writing but may need some changes and editing therefore comments are as follows; 

-Abstract needs brush up in English and also the result not well mentioned

-Keywords are more than 5 words and some words repeat like water quality and is better some words be different from the title.

- What is the main impute of this research? And what was different by previous work please clearly mentioned in the introduction.

- What was the time period of the Landsat and Sentinel data? Was same period for both lake?

- How did you apply the satellite reference data in the model? Which month? And how you do the accuracy assessment?

- Graphs can be modify by better line and color 

- Conclusion not well support the result.

- Minor English grammar check may apply.

- Minor changes and grammar applicable  

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Review report of "Assessment of Machine Learning Models for Remote Sensing of Water Quality in Lakes Cajititlán and Zapotlán, Jalisco - Mexico."

General Opinion: The manuscript evaluated a machine learning technique to monitor lake water quality parameters such as chlorophyll-a, turbidity, and total suspended solids (obtained from remote sensing (e.g., Landsat-8 and Sentinel-2) and in situ data). The study was conducted in Cajititlán and Zapotlán in Jalisco, Mexico. Water quality parameters obtained from ground-based (i.e., from the Mexican National Water Quality Monitoring Network) and remote-sensed-based datasets (e.g., Landsat-8 and Sentinel-2) were used for correlation analysis. The study reported different coefficient of determination (r²) values for the lakes under study.

The manuscript is original, and the analysis contributes to the literature. The result findings may serve as input for other resource management operational decisions. However, few weaknesses were observed. I require the authors to address them scientifically. To guide the authors, I have concerns; I have itemized my comments and suggestions in the attached file.

Comments for author File: Comments.pdf

Minor editing of English language required.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Authors used landsat and sentinel dataset to retrieve water quality. Generally, this manuscript used machine learning models to construct retrieved models to validate the accuracy of water quality. Some comments might be useful for authors.

The structure of the abstract is unreasonable or the elements in the abstract are incomplete. A good abstract contains four elements: background, methods, results, and conclusions. The abstract had better been revised according to the following structural arrangement: background and aim, methods, results or findings, conclusions or significance.

Introduction: introduction is important, however, authors talk less about the methods in existing studies, suggest authors to add the description of methods part, what about other similar studies, how does they do this job.

Samples part, authors collects several samples to as the situ samples, however, machine learning needs many samples to get good models, is the samples number meet the requirement? this is important.

The structure of the Discussion part of this article is incomplete. The section of Results represents the heart of a paper, and the section of Discussion is the paper’s nerve center. The section of Discussion in a paper is generally involved with 3 or 4 parts: (1) main points, which response to the questions put in introduction; (2) comments on related studies or problems; (3) shortcomings or deficiency in study method or process; (4) conclusions, which can be separated to make the final section.

After these modification, this manucript could be further consideration.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

Reviewer comments

The authors have done a good job. Their novelty lies in the development of machine learning (ML) models to estimate 3 water quality variables (chl-a, turbidity and suspended solids) in two inland aquatic ecosystems in Mexico.

However, there are some points that I suggest incorporating.

Point 1. Line 24, 25. Therefore, in recent years, continuous 24 monitoring campaigns for WQ parameters have increased in several Latin American 25 countries. This sentence needs references.

Point 2. Please rewrite this sentence. The purpose is to understand and prevent WQ deterioration through the analysis of data resulting from monitoring campaigns.

Point 3. I suggest including some current references on the estimation of water quality parameters in Latin America (poor countries). For example:

Chile(https://www.mdpi.com/2072-4292/14/18/4568)

Brasil (https://www.mdpi.com/2073-4441/14/3/400)

Argentina (https://www.sciencedirect.com/science/article/abs/pii/S0895981119302226)

Point 4. Only r2 is highlighted in the abstract but RMSE and MAE errors are not included, why?

Point 5. Is it possible to present fig, 7 and 8 as 1 figure? For example:

fig. 7 Evaluation of the predictive ability of the ML models for (a) Lake Cajititlán and (b) Lake Zapotlán. The models are distinguished by symbols and solid lines in three shades of blue for each WQ parameter. Landsat-8 (L8), Sentinel-2 (S2).

It may be possible to integrate the discussion information from both figures. is too long.

Point 6. Lines 300-309. Why was the MAE not included in the discussion? To visualize the differences please place the same scale in figure 9. what does the dashed line mean?

Other suggestions

use the same format of coordinates in the maps, fig 1 does not have the same format as figs. 2,3 etc.

move fig. 2 below section 2.4.1

line 32. Missing ref?

I suggest you check the English

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Authors have addressed all concerns. Well done.

Reviewer 4 Report

The authors of the manuscript Assessment of Machine Learning Models for Remote Sensing of Water Quality in Lakes Cajititlán and Zapotlán, Jalisco - Mexico have incorporated and responded the suggestions I previously sent. So, I consider that it is possible to publish their paper.  

Best

L. G-R

Minor editing of English language required