Spatiotemporal Monitoring of Cyanobacterial Blooms and Aquatic Vegetation in Jiangsu Province Using AI Earth Platform and Sentinel-2 MSI Data (2019–2024)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The study employs Sentinel-2 MSI data and the AI Earth platform to develop a monitoring method aimed at detecting cyanobacterial blooms and aquatic vegetation in small and medium-sized lakes and reservoirs. The proposed system effectively demonstrates ecosystem responses to environmental changes and restoration efforts. The research covers multiple lakes and identifies notable spatiotemporal variations, offering potentially valuable tools for aquatic ecosystem monitoring. Nonetheless, several points require clarification from the authors.

1. Discuss how the monitoring system can be used in other regions or lake ecosystems and what its limitations are.
2. Ensure consistent use of terminology and classification standards throughout the text to maintain clarity and coherence. For example, choose either Cyanobacterial Blooms or Cyanobacterial Bloom, and either Aquatic Vegetation or Aquatic Plant, and use the selected terms consistently throughout the manuscript.
3. The paper discusses cyanobacterial blooms and aquatic vegetation separately but does not provide a quantitative analysis of their interactions.
4. Include recommendations for future ecological management, such as establishing an early warning system for cyanobacterial blooms.
5. Maintain consistency in the capitalization of words in subheadings throughout the manuscript.
6. Conduct a comprehensive review of the manuscript to identify and correct errors, such as the typo in line 156, where 'Satellite a data' should be corrected to 'Satellite data.
7. The section on future research is too limited. Add suggestions on data integration and algorithm improvements.
8. In Figure 1, include the area and average depth of each lake to better show the study area's characteristics.
9. Check that subscripts and superscripts in formulas are formatted the same way. For example, in Formula 10, make sure R_(i_Aqu_vegetation) and R_(i_Aqu_plant) are consistent.
10. The word "Long-Term" in the title is not correct because the study period is only from 2019 to 2024. It is better to take it out.

Author Response

The study employs Sentinel-2 MSI data and the AI Earth platform to develop a monitoring method aimed at detecting cyanobacterial blooms and aquatic vegetation in small and medium-sized lakes and reservoirs. The proposed system effectively demonstrates ecosystem responses to environmental changes and restoration efforts. The research covers multiple lakes and identifies notable spatiotemporal variations, offering potentially valuable tools for aquatic ecosystem monitoring.

Thank you for your valuable comments and feedback. We are pleased that you recognize the approach employed in our study. This research utilizes Sentinel-2 MSI data and the AI Earth platform to develop a monitoring method aimed at detecting cyanobacterial blooms and aquatic vegetation in small- and medium-sized lakes and reservoirs. The proposed system effectively demonstrates ecosystem responses to environmental changes and restoration efforts.

The study covers multiple lakes and identifies notable spatiotemporal variations, offering potentially valuable tools for aquatic ecosystem monitoring. We will continue to optimize the application of this method and look forward to contributing more to ecological restoration and environmental management in future research.

 

Point 1: Discuss how the monitoring system can be used in other regions or lake ecosystems and what its limitations are.

Response 1: Thank you for your insightful comment. We appreciate your suggestion to discuss how the monitoring system can be applied to other regions or lake ecosystems, as well as its limitations.

In response to your feedback, we have expanded the manuscript to include a discussion on the potential applications of the monitoring system in different regions and lake ecosystems. The system, based on Sentinel-2 MSI imagery and the AI Earth platform, can be adapted to monitor various freshwater ecosystems worldwide, provided that the lakes share similar characteristics, such as nutrient enrichment, water turbidity, and the occurrence of cyanobacterial blooms. The system's scalability allows it to be deployed in various geographical locations with sufficient satellite data coverage and temporal resolution.

However, the system has limitations in broader applications. It is specifically designed for detecting cyanobacterial blooms in freshwater ecosystems and may not be suitable for monitoring other types of marine algae, such as red tides caused by dinoflagellates. Similarly, diatom blooms in certain freshwater lakes may not be detected accurately due to their distinct spectral properties compared to cyanobacteria. Additionally, the system may face challenges in regions with frequent cloud cover or insufficient satellite coverage, which could affect the accuracy of the monitoring.

 

Point 2: Ensure consistent use of terminology and classification standards throughout the text to maintain clarity and coherence. For example, choose either Cyanobacterial Blooms or Cyanobacterial Bloom, and either Aquatic Vegetation or Aquatic Plant, and use the selected terms consistently throughout the manuscript.

Response 2: Thank you for your valuable feedback. We appreciate your suggestion to ensure consistent use of terminology and classification standards throughout the manuscript. In response to your comment, we have carefully reviewed the entire text and made the necessary adjustments. We have standardized the terms used, choosing "Cyanobacterial Bloom" and "Aquatic Vegetation" consistently throughout the manuscript. This adjustment helps improve clarity and coherence, as well as maintains consistency in the scientific language used.

 

Point 3: The paper discusses cyanobacterial blooms and aquatic vegetation separately but does not provide a quantitative analysis of their interactions.

Response 3: Thank you for your valuable feedback. In response to your comment that the paper discusses cyanobacterial blooms and aquatic vegetation separately but does not provide a quantitative analysis of their interactions, we have made corresponding adjustments in the discussion and conclusion sections. In these sections, we have added relevant quantitative data to better describe the changes in both cyanobacterial blooms and aquatic vegetation, providing more quantitative support for the study. Thank you for your suggestion, and we believe these additions will enhance the completeness of the paper.

 

Point 4: Include recommendations for future ecological management, such as establishing an early warning system for cyanobacterial blooms.

Response 4: Thank you for your valuable feedback. In response to your suggestion to include recommendations for future ecological management, such as establishing an early warning system for cyanobacterial blooms, we have added the following content in the Future Outlook section:

Future ecological management strategies should prioritize the development of a more efficient and timely early warning system for cyanobacterial blooms. Currently, satellite data processing via cloud-based platforms is subject to a delay of approximately 10 hours. To enhance the timeliness of bloom forecasts, we recommend implementing a localized automated processing system that utilizes API-based satellite data downloads. By processing the data locally, response times could be reduced to 1–2 hours, facilitating more rapid and precise early warnings. This advancement would enable stakeholders to take prompt, informed action to mitigate the impact of blooms.

 

Point 5: Maintain consistency in the capitalization of words in subheadings throughout the manuscript.

Response 5: Thank you for your valuable feedback. We have made the necessary adjustments to ensure consistency in the capitalization of words in subheadings throughout the manuscript. We appreciate your suggestion, which has helped improve the overall consistency and formatting of the paper.

 

Point 6: Conduct a comprehensive review of the manuscript to identify and correct errors, such as the typo in line 156, where 'Satellite a data' should be corrected to 'Satellite data.

Response 5: Thank you for your valuable feedback. We have conducted a comprehensive review of the manuscript and corrected all errors, including the typo in line 156 where "Satellite a data" has been corrected to "Satellite data." We appreciate your suggestion, which has helped improve the accuracy and consistency of the paper.

 

Point 7: The section on future research is too limited. Add suggestions on data integration and algorithm improvements.

Response 7: Thank you for your valuable feedback. In response to your suggestion that the future research section is too limited, we have added recommendations on data integration and algorithm improvements in the Future Outlook section. Specifically, we suggest:

Enhancing Data Integration: Future research should focus on improving the integration of multi-source data, combining satellite imagery, water quality sensors, meteorological data, and other environmental variables to gain a more comprehensive understanding of the factors driving bloom dynamics.

Algorithm Improvements: We recommend developing more advanced algorithms, such as machine learning models, to optimize the integration of these diverse datasets and improve the accuracy of bloom predictions. Additionally, refining detection algorithms for smaller-scale blooms and distinguishing between different cyanobacterial species could lead to more precise and tailored management strategies.

These additions aim to deepen future research and further enhance the scientific and practical aspects of ecological management. Thank you again for your valuable suggestion.

 

Point 8: In Figure 1, include the area and average depth of each lake to better show the study area's characteristics.

Response 8: Thank you for your valuable feedback. In response to your suggestion, we have made the necessary adjustments to Figure 1 by including the area and average depth of each lake. This will help better illustrate the characteristics of the study area. We appreciate your input, which has enhanced the clarity and detail of the figure.

 

Point 9: Check that subscripts and superscripts in formulas are formatted the same way. For example, in Formula 10, make sure R_(i_Aqu_vegetation) and R_(i_Aqu_plant) are consistent.s.

Response 9: Thank you for your valuable feedback. In response to your suggestion, we have carefully reviewed the formulas in the manuscript to ensure that all subscripts and superscripts are consistently formatted.

 

Point 10:  The word "Long-Term" in the title is not correct because the study period is only from 2019 to 2024. It is better to take it out.

Response 10: Thank you for your valuable feedback. We understand your concern regarding the use of "Long-Term" in the title, given that the study period spans from 2019 to 2024. In response to your suggestion, we have removed "Long-Term" from the title for greater accuracy. The revised title now reads: "Spatiotemporal Monitoring of Cyanobacterial Blooms and Aquatic Vegetation in Jiangsu Province Using AI Earth Platform and Sentinel-2 MSI Data (2019-2024) ." We appreciate your suggestion, which has helped improve the clarity and appropriateness of the title.

Reviewer 2 Report

Comments and Suggestions for Authors

Review of Xie et al. RS 3666157 “Long-Term Spatiotemporal Monitoring of Cyanobacterial 2 Blooms and Aquatic Vegetation in Jiangsu Province Using AI 3 Earth Platform and Sentinel-2 MSI Data”

 

This study capitalizes on a time series of Sentinel-2 MSI imagery, downloaded from the AI Earth (AIE) platform developed by Alibaba 17 DAMO Academy to separate, and thus be able to monitor, cyanobacterial blooms and aquatic vegetation in multiple lakes along a Norht-South gradaient in Jiangsu Province, China over multiple years between 2019-2024.

I will admit, at first I was very frustrated with this study because there is some explanation lacking in the methods and the results are not synthesized to highlight the important contributions of the study. However, once I finally got to the discussion, the authors show they do understand the ecology and the issues that may impact their results.  This paper would benefit from a major regrouping of the figures to exactly illustrate the points in the discussion.  Yes, this will tak ea bit of work to re-roganize, cut and determine what goes to supplementary material but if the authors do this, the paper will be solid. 

 

The comments written below the line were generated as I was reading the paper from start to finish AFTER I looked through the figures and got a bit lost.  Please consider how you will use them to update the materials with details that should be included in the paper.

 

A couple of other important concerns are pesented here:

• Regarding the AIE-DAMO platform. This paper makes it seem like this is a platform that delivers “clean”/”processed” imagery from multiple sources to the “public”.  I do not find any links for public access or documentation.  The reference provided “Xu H., Man, Y., Yang, M., Wu, J., M., Zhang, Q., & Wang, J. Analytical Insight of Earth: A Cloud-Platform of Intelligent Com-557 puting for Geospatial Big Data. Multiagent and Grid Systems, 2023, arXiv:2312.16385. DOI: 558 https://doi.org/10.48550/arXiv.2312.16385. “is not a peer-reviewed text or accessible project.

 

• Steps performed in the AI Platform versus additional processes following download need to be clarified in the methods and Figure 2 needs to show what steps are pre-processing by AIE-DAMO and what is being done by the researcher.

 

• IF the “AI Earth Platform” is not publicly available and only used for data download, this should be removed from the title because it leads one to believe there is novel invention in the research.

_______________________________________________

This is a fine study in how to investigate multiple remote sensing indices from Sentinel-2 MSI for monitoring different surface aquatic parameters, but the paper leaves me with a feeling that there is a lack of understanding of the ecology of the system and radiative physics of remote sensing in the writing that would convince me this is a strong scientific paper.  I realize the Remote Sensing (the journal) is more strongly focused on the technical side, which is acceptable, but I would like to see writing that expresses a more holistic understanding.

 

At the very least, in the introduction and methods, please update explanations of topics and the references they refer to, to reflect the RS platforms and/or the specific ecological components that underpin.  I have added notes in the document and provided some examples below.

 

Specifics:

There is no detail about how or why additional indices were chosen and the references provided are not adequate as they only describe two indices.  Yes, I believe the introductions to the referenced papers explain other RS indices but there is a lack of direct reference to what is actually being measured in this study.  For example: there are multiple components to aquatic cyanobacteria.  There is no discussion of which of these is optically relevant for satellite remote sensing.  In addition, there is a mixing of references to remote sensing platforms of different scales (UAV data is not Sentinel).  This is acceptable if the details of the differences in terms of radiometric impacts are explained.  Simply referencing indices because they may have been used in a previous study is not acceptable.

 

 

All the acronyms in Figure 2 diagram need to be explained in the figure caption.  If an index – the citation needs to be provided, i.e. What are FAI, VFI and where were they derived?

 

Line 143-145. This is referring to a UAV study.  Yes, RS but when the data are from satellites, the underlying information needs to be provided at the same scale or acknowledge a different and explain how the radiative physics maybe the same but there are a multitude of factors that impact the signal as one scales up from drones to satellites.

 

Equations line 150. These index names need to be more specific if proposing new indices based on a specific instrument (MSI) and ecological parameter (Cyanobacterial blooms).  Are these the VFI in the Figure 2 processing step?

 

FIgure 3.

• This figure needs A, B, C, D added to the imagery boxes and a clear explanation in this figure legend as to what each section represents and where the legend is being applied. For example, the upper two boxes, black is water but in the bottom boxes it is supposed to be a category of "clouds or unknown".  It is then confusing to have white clouds in the mapped data. 

 

• The bottom boxes should just be the prediction output, so one can actually see that the aquatic vegetation is mostly concentrated around the shorelines or islands - or what I am interpreting as islands.

 

• What are the gradations of greenish-black that show up in the CIR image of "South of Lake Taihu". Is this turbidity?

 

• It does not seem like turbidity is being accounted for thin this study.

 

 

Figure 5: This figure is what I would call a data dump.  The organization of the sites is not explained in the legend nor are any of the parameters. Because all this information is being stuffed into one figure, one canʻt read the axes to try to interpret what is going on. 

 

• Then there are little things, like it is impossible to see the line that has been placed over the bars in the "I" parts of the graphs.

 

• The scale and compass rose relating to the maps should be under the maps, not at the bottom of the figure.

 

• One should be able to look at only the figures of a paper and understand the main story.

 

• If the signal is in the 6 lakes in the South versus the north, then show the bloom frequency first and then dig into the rest of the cover differences etc.

 

 

YES - Finally!!!! Sediment AND this likely impacts the RS in multiple ways.  Submerged sediments likely have cyanobacteria emerging.  They can be mobile in the water column and precursors to a full cyanobacteria bloom.  This is not accounted for in this study.  Sediment is not mentioned until here, line 320.

 

Recommendation: Accept with Major Revisions: Overall, if the authors can “illustrate” their clearly and simply results in their graphs which consists of creating summary graphs and using online supplementary materials to supply supporting information based on the main points laid out in their disucssion and providing additional details in the methods as well as their figure captions, then this paper could be recommended for publication.  In its current form, there are problems with referenced materials and the figures are what in an advisorʻs term are a “data dump” which does not support the advancement of remote sensing science.

Author Response

Point 1: Regarding the AIE-DAMO platform. This paper makes it seem like this is a platform that delivers “clean”/”processed” imagery from multiple sources to the “public”.  I do not find any links for public access or documentation.  The reference provided “Xu H., Man, Y., Yang, M., Wu, J., M., Zhang, Q., & Wang, J. Analytical Insight of Earth: A Cloud-Platform of Intelligent Com-557 puting for Geospatial Big Data. Multiagent and Grid Systems, 2023, arXiv:2312.16385. DOI: 558 https://doi.org/10.48550/arXiv.2312.16385. “is not a peer-reviewed text or accessible project.

Response 1: Thank you for your valuable comment regarding the accessibility and documentation of the AI Earth (AIE) platform. We clarify that the AIE platform, developed by Alibaba DAMO Academy, is fully publicly accessible at https://aiearth.aliyun.com. It offers a wide array of remote sensing datasets and analytical tools for geospatial analysis. While the platform is open to the public, we acknowledge that peer-reviewed documentation remains limited. We have revised the manuscript to reflect this clarification and included the direct link for readers interested in exploring the platform's capabilities.

Furthermore, our research team has developed several online remote sensing analysis tools for environmental monitoring using the AIE platform in combination with Streamlit. These systems are available at http://www.zgwxsj.com. Specifically, the “Lake Cyanobacterial Blooms and Aquatic Vegetation Monitoring” module, directly relevant to this study, can be accessed at http://www.zgwxsj.com:8509. This platform enables users to interactively explore time-series monitoring results based on the methodology presented in this paper.

Additionally, we have deployed a parallel application with similar functionality on the Google Earth Engine (GEE) platform. This version is publicly accessible at https://songting1207.users.earthengine.app/view/s2bve, offering an alternative interface for users to visualize the monitoring outputs. Both systems are documented in the Code Availability section of the revised manuscript.

 

Point 2: Steps performed in the AI Platform versus additional processes following download need to be clarified in the methods and Figure 2 needs to show what steps are pre-processing by AIE-DAMO and what is being done by the researcher.

Response 2: We thank the reviewer for the insightful comment. In the revised manuscript, Section 2.2 explicitly outlines the division of processing tasks between the AI Earth (AIE) platform and our research team. We clarify that the AIE platform provides Sentinel-2 Level-2A imagery, consisting of atmospherically corrected surface reflectance data, which serves as the foundational input for our analysis.

In addition to data access and atmospheric correction, the AIE platform—comparable to Google Earth Engine (GEE)—offers multi-source satellite datasets, a cloud-based development environment, and a comprehensive suite of analytical tools. Using this platform, our team implemented the complete analytical workflow, including data acquisition, spatial masking, time-series construction, algorithm development, and result visualization. This fully integrated, end-to-end process ensured efficient and reproducible generation of remote sensing products. The specific implementation and logic of these steps are further described in Section 2.3 of the revised manuscript. To improve transparency, Figure 2 has been revised to clearly distinguish processing steps performed within the AIE platform from those carried out externally.

 

Point 3: IF the “AI Earth Platform” is not publicly available and only used for data download, this should be removed from the title because it leads one to believe there is novel invention in the research.

Response 3: We appreciate the reviewer’s concern. The AI Earth (AIE) platform, developed by Alibaba DAMO Academy, is fully publicly accessible via [https://aiearth.aliyun.com](https://aiearth.aliyun.com). It is not merely a data download portal but a comprehensive, cloud-based geospatial computing environment. In this study, AIE was used not only to access Sentinel-2 Level-2A products, but also to perform key pre-processing tasks—including spatial masking, temporal aggregation, and spectral index calculation—using its integrated analysis modules.

Given AIE’s central role in supporting scalable, cloud-based remote sensing workflows—comparable in functionality to platforms like Google Earth Engine (GEE)—we consider its inclusion in the title both appropriate and justified. Nevertheless, we have revised the manuscript to more clearly describe the platform’s capabilities and public accessibility in the Methods and Code Availability sections.

 

Point 4: There is no detail about how or why additional indices were chosen and the references provided are not adequate as they only describe two indices.  Yes, I believe the introductions to the referenced papers explain other RS indices but there is a lack of direct reference to what is actually being measured in this study.  For example: there are multiple components to aquatic cyanobacteria.  There is no discussion of which of these is optically relevant for satellite remote sensing..

Response 4: We sincerely thank the reviewer for raising this important point. In the revised manuscript, we have substantially clarified the rationale behind the selection of spectral indices and explicitly identified the optically relevant components of aquatic cyanobacteria targeted in this study.

Specifically, our detection strategy focuses on floating cyanobacterial blooms, which are primarily composed of surface-accumulated colonies of Microcystis spp. and other buoyant genera. The optically active components relevant to satellite remote sensing include phycocyanin, chlorophyll-a, and associated pigments, as well as the cellular structure and gas vacuoles that enhance reflectance in the red-edge and near-infrared regions. These spectral features informed our use of indices such as the Floating Algae Index (FAI), red-edge-based band combinations, and novel normalized difference formulations designed to suppress thin cloud and turbidity interference.

We have expanded Section 2.3.2 to explain the design and selection of the three newly developed indices, which build upon these optical principles and are specifically tailored to the Sentinel-2 MSI sensor's spectral configuration.

We appreciate the reviewer’s observation that the two initially cited references did not fully cover all indices used, and we have now revised both the text and citations to more accurately reflect the methodological foundation of our remote sensing approach.

 

Point 5: In addition, there is a mixing of references to remote sensing platforms of different scales (UAV data is not Sentinel).  This is acceptable if the details of the differences in terms of radiometric impacts are explained. Simply referencing indices because they may have been used in a previous study is not acceptable.

Response 5: We appreciate the reviewer’s insightful comment. The previously cited reference on UAV-based remote sensing data was not suitable, as it did not align with the focus of this study, which is on satellite-based monitoring of cyanobacterial blooms. We have replaced it with Fang et al. (2018), which directly addresses the limitations of the Floating Algae Index (FAI) under thin cloud conditions. This reference provides a detailed explanation of how the Adjusted FAI (AFAI) method was developed to reduce false positives caused by cloud interference, particularly from thin clouds. The updated citation is:

Fang, C., Song, K. S., Shang, Y. X., Ma, J. H., Wen, Z. D., & Du, J. (2018). Remote sensing of harmful algal blooms variability for Lake Hulun using adjusted FAI (AFAI) algorithm. Journal of Environmental Informatics, 34(2), 108-122. https://doi.org/10.3808/jei.201700385

We believe this revision more accurately reflects the methodology employed in this study and provides clearer insights into the challenges of using FAI under these conditions.

 

Point 6: All the acronyms in Figure 2 diagram need to be explained in the figure caption.  If an index – the citation needs to be provided, i.e. What are FAI, VFI and where were they derived?

Response 6: We appreciate the reviewer’s thoughtful comment. In the revised manuscript, we have updated the figure caption to clarify all acronyms used in Figure 2 and included the appropriate citations for the indices.

 

Point 7: Line 143-145. This is referring to a UAV study. Yes, RS but when the data are from satellites, the underlying information needs to be provided at the same scale or acknowledge a different and explain how the radiative physics maybe the same but there are a multitude of factors that impact the signal as one scales up from drones to satellites.

Response 7: We thank the reviewer for this valuable comment. The response to this point is the same as our reply to point 5.

 

Point 8: Equations line 150. These index names need to be more specific if proposing new indices based on a specific instrument (MSI) and ecological parameter (Cyanobacterial blooms).  Are these the VFI in the Figure 2 processing step?

Response 8: Thank you for this thoughtful comment. In the revised manuscript, we have clarified that the three indices presented in line 150 are newly developed and specifically designed for cyanobacterial bloom detection using Sentinel-2 MSI Level-2A data under conditions of residual cloud and shadow contamination. These indices are not the same as the VFI (Vegetation Frequency Index) shown in the Figure 2 processing step. VFI was developed in this study to distinguish between frequent bloom regions and stable aquatic vegetation zones, based on long-term NDVI time series.

To avoid confusion and improve clarity, we have renamed the three indices as follows:

CBI1 – Cyanobacterial Bloom Index 1

CBI2 – Cyanobacterial Bloom Index 2

CBI3 – Cyanobacterial Bloom Index 3

These names now appear consistently throughout the manuscript. The purpose and derivation of CBI indices have also been elaborated in the Methods section, and their distinction from VFI is now clearly stated in both the figure and the corresponding text.

 

Point 9: FIgure 3. (1) This figure needs A, B, C, D added to the imagery boxes and a clear explanation in this figure legend as to what each section represents and where the legend is being applied. For example, the upper two boxes, black is water but in the bottom boxes it is supposed to be a category of "clouds or unknown". It is then confusing to have white clouds in the mapped data. (2) The bottom boxes should just be the prediction output, so one can actually see that the aquatic vegetation is mostly concentrated around the shorelines or islands - or what I am interpreting as islands. (3) What are the gradations of greenish-black that show up in the CIR image of "South of Lake Taihu". Is this turbidity? (4) It does not seem like turbidity is being accounted for thin this study.

Response 9: We thank the reviewer for the detailed suggestions regarding Figure 3. In the revised version, we have implemented the following improvements:

• Panel Labels and Legend Clarification: Alphabetical labels (A, B, C, D) have been added to all subfigures in Figure 3. The figure caption has been revised to clearly explain the content of each panel and to explicitly state that the legend applies only to the classification results (B and D).
• Adjustment of Bottom Panels: As recommended, the lower panels (C and D) now display only the prediction outputs, enabling clearer visualization of the spatial distribution of aquatic vegetation, particularly its concentration along shorelines and around islands.
• Greenish-Black Gradients in CIR Imagery: The reviewer is correct in observing greenish-black gradients in the false-color (CIR) image of the southern part of Lake Taihu. These gradients correspond to turbidity levels, with darker tones indicating lower turbidity.
• Consideration of Turbidity: The influence of turbidity on the classification process has been accounted for. A detailed description of how turbidity effects were handled is provided in Section 2.3.3.

 

Point 10: FIgure 5. (1) This figure is what I would call a data dump. The organization of the sites is not explained in the legend nor are any of the parameters. Because all this information is being stuffed into one figure, one canʻt read the axes to try to interpret what is going on. (2) Then there are little things, like it is impossible to see the line that has been placed over the bars in the "I" parts of the graphs. (3)The scale and compass rose relating to the maps should be under the maps, not at the bottom of the figure. (4) One should be able to look at only the figures of a paper and understand the main story. (5) If the signal is in the 6 lakes in the South versus the north, then show the bloom frequency first and then dig into the rest of the cover differences etc.

Response 10: We sincerely thank the reviewer for the detailed feedback on Figure 3. In response, we have carefully revised the figure and its accompanying description to address all the concerns raised:

• Figure organization and clarity of parameters: We agree that the original version of the figure was overly dense and lacked sufficient structural clarity. To improve interpretability, we have divided the content into two separate figures—Figure 5 for lakes north of the Yangtze River and Figure 6 for those in the south. The revised figure captions clearly explain the spatial grouping of sites (Panels A–F and G–L), define the parameters shown in subpanels I and II, and specify the classification categories used. These changes significantly enhance the clarity, flow, and readability of the figures.
• Line visibility in subpanel I: We have revised the visual design of the trend line overlaid in subpanel I, increasing its thickness and contrast to ensure that it is easily distinguishable above the bars. This adjustment improves visual clarity while maintaining the integrity of the data.
• Placement of scale bar and compass rose: As suggested, we have corrected the map layout by embedding the scale bar and compass rose directly within each individual map panel, rather than placing them collectively at the bottom of the figure. This revision provides clear spatial context and adheres to cartographic conventions.
• Figures as standalone storytelling elements: We fully agree that figures should be interpretable without relying entirely on the main text. In the revised version, the figures now follow a logical progression from spatial bloom frequency patterns to quantitative summaries of classification areas and trends. The expanded figure captions further improve self-containment and accessibility.
• Highlighting North–South contrast in bloom frequency: In accordance with the reviewer’s recommendation, we have adjusted the order of presentation to first emphasize the spatial contrast in bloom frequency between northern and southern lakes, followed by subsequent details on aquatic vegetation dynamics. This new structure brings the geographic signal to the forefront and better aligns with the manuscript’s core narrative.

 

Point 11: Sediment AND this likely impacts the RS in multiple ways. Submerged sediments likely have cyanobacteria emerging.  They can be mobile in the water column and precursors to a full cyanobacteria bloom.  This is not accounted for in this study.  Sediment is not mentioned until here, line 320.

Response 11: We thank the reviewer for this important comment. We agree that submerged sediments and their interaction with cyanobacteria can influence remote sensing signals in multiple ways, including the potential emergence of cyanobacteria from sediment and their movement within the water column. While our study primarily focuses on surface-floating cyanobacterial blooms, we acknowledge the broader ecological processes underlying bloom dynamics.

As noted in our previous response to Point 9(4), the influence of turbidity—often linked to suspended sediments—has been explicitly considered during image classification. Section 2.3.3 of the revised manuscript provides a detailed explanation of how turbidity was accounted for, including the use of indices designed to minimize misclassification caused by turbid water and bottom reflectance.

Although the role of sediment as a biological precursor to bloom initiation is not the focus of this remote sensing-based study, we have added a clarification in the Discussion section to acknowledge this limitation and highlight it as an important avenue for future research.

Reviewer 3 Report

Comments and Suggestions for Authors

Review of the Paper "Long-Term Spatiotemporal Monitoring of Cyanobacterial Blooms and Aquatic Vegetation in Jiangsu Province Using AI Earth Platform and Sentinel-2 MSI Data"

 

This paper presents an automated monitoring system for tracking cyanobacterial blooms and aquatic vegetation in Jiangsu Province using Sentinel-2 MSI data and the AI Earth (AIE) cloud platform. It integrates phenology-based algorithms to distinguish cyanobacterial blooms from aquatic vegetation, leveraging temporal persistence (Vegetation Frequency Index) and spectral indices (e.g., FAI, NDVI). The study spans 2019–2024, revealing a north-south gradient in bloom intensity (higher in southern lakes) and divergent vegetation trends linked to ecological restoration and fishing bans. Key contributions include the use of high-resolution Sentinel-2 data for small/medium lakes and cloud-based processing for scalability.

 

1. The presentation effect of the figures is too poor. It is recommended to make significant modifications and increase clarity. It is suggested to use professional software for drawing. Especially Figure 5 and Figure 6.
2. The full text of the abstract only uses qualitative analysis, such as’ high ‘, for discussion. It is recommended to revise it to a quantitative analysis to enhance the readability of the abstract.
3. Regarding the accuracy verification of the model, it was not clearly discussed in the method section, but directly appeared in the first section of the results. It is recommended to make modifications.
4. The 'rs' in 𝑅𝑟𝑠 is generally used as a subscript, it is recommended to revise it.
5. The depth of discussion on the differences between the North and South is insufficient. It is suggested to further enhance the contribution and significance of this article by combining climate and hydrological factors, social factors, etc.
6. The conclusion mainly involves qualitative analysis, lacking quantitative expression, and needs to highlight the contribution of this article.

Comments on the Quality of English Language

Does not affect reading

Author Response

This paper presents an automated monitoring system for tracking cyanobacterial blooms and aquatic vegetation in Jiangsu Province using Sentinel-2 MSI data and the AI Earth (AIE) cloud platform. It integrates phenology-based algorithms to distinguish cyanobacterial blooms from aquatic vegetation, leveraging temporal persistence (Vegetation Frequency Index) and spectral indices (e.g., FAI, NDVI). The study spans 2019–2024, revealing a north-south gradient in bloom intensity (higher in southern lakes) and divergent vegetation trends linked to ecological restoration and fishing bans. Key contributions include the use of high-resolution Sentinel-2 data for small/medium lakes and cloud-based processing for scalability.

We sincerely thank the reviewer for the thoughtful and accurate summary of our work.

We are pleased that the reviewer has recognized the key contributions of our study, including the application of high-resolution Sentinel-2 MSI data, the integration of phenology-based algorithms—such as the Vegetation Frequency Index—and spectral indices (e.g., FAI, NDVI) for effectively distinguishing cyanobacterial blooms from aquatic vegetation. We also appreciate the acknowledgement of our cloud-based monitoring system implemented on the AI Earth (AIE) platform, which enables scalable, automated analysis. Additionally, we are grateful for the recognition of our spatiotemporal findings, notably the observed north–south gradient in bloom intensity and the divergent vegetation dynamics driven by ecological restoration and fishing bans.

Looking ahead, we plan to further improve the system’s adaptability and generalizability across a wider range of climatic conditions and watershed environments.

We thank the reviewer again for their constructive evaluation and kind support.

 

Point 1: The presentation effect of the figures is too poor. It is recommended to make significant modifications and increase clarity. It is suggested to use professional software for drawing. Especially Figure 5 and Figure 6.

Response 1: Thank you for your valuable feedback regarding the presentation quality of the figures. We agree that clear and professional visualizations are essential for effectively conveying the results of our study.

In response to your suggestion, we have made significant improvements to the figures, enhancing their clarity and overall presentation.

 

Point 2: The full text of the abstract only uses qualitative analysis, such as’ high ‘, for discussion. It is recommended to revise it to a quantitative analysis to enhance the readability of the abstract.

Response 2: Thank you for your valuable feedback. We have revised the abstract to include more quantitative analysis, replacing qualitative terms with specific numerical data. This revision enhances the clarity and readability of the abstract.

 

Point 3: Regarding the accuracy verification of the model, it was not clearly discussed in the method section, but directly appeared in the first section of the results. It is recommended to make modifications.

Response 3: Thank you for your insightful comment regarding the accuracy verification of the model. We appreciate your suggestion to clarify this aspect in the method section rather than introducing it directly in the results section. In response to your feedback, we have made the necessary adjustments and now provide a detailed discussion of the accuracy verification in the method section. This modification ensures that the methodology is clearly explained before the results are presented.

 

Point 4: The 'rs' in ??? is generally used as a subscript, it is recommended to revise it .

Response 4: Thank you for your valuable feedback. We have revised the expression as per your recommendation, using "rs" as a subscript in ??? to align with standard notation.

 

Point 5: The depth of discussion on the differences between the North and South is insufficient. It is suggested to further enhance the contribution and significance of this article by combining climate and hydrological factors, social factors, etc.

Response 5: Thank you for your valuable suggestion. We have taken your feedback into account and have significantly enhanced the discussion on the differences between the North and South in Section 4.2. In this revision, we have further integrated the influence of climate and hydrological factors, as well as social factors such as agricultural activities, urbanization, and fishing pressures. These factors are now discussed in greater depth to highlight their contribution to regional differences in cyanobacterial bloom dynamics. We believe that this strengthened discussion improves the contribution and significance of the article.

 

Point 6: The conclusion mainly involves qualitative analysis, lacking quantitative expression, and needs to highlight the contribution of this article.

Response 6: Thank you for your valuable feedback. We appreciate your suggestion to incorporate more quantitative analysis and to emphasize the contributions of this study.

In response to your comment, we have revised the conclusion to include quantitative data that better highlights the study's findings and contributions. Additionally, we have made an effort to more clearly outline the significance of the developed remote sensing monitoring system and its advancements in both ecological monitoring and water quality management.

The revised conclusion now emphasizes the quantitative results observed, such as the specific changes in bloom intensity and aquatic vegetation coverage, as well as the percentage increases in affected areas. 

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

First of all, thank you very much for the author's revisions.

The quality of some images still cannot be recognized and labeled, such as Figures 5/6/8/9.

The captions for Figures 4 and 7 require information such as circular error lines to enhance readability.

Comments on the Quality of English Language

Does not affect reading.

Author Response

Figures 4 and 7 have been revised accordingly. In addition, all high-resolution images, including the updated versions of Figures 4 and 7, have been bundled and uploaded for your review.
Thank you very much for your valuable guidance and support throughout the revision process.