Fluoride Enrichment and Health Risks in the Aksu River Basin Oasis: Implications for Soil–Groundwater Systems

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Below, I am sending some suggestions regarding the article "Unveiling the Hidden Threat: Fluoride Risks in the Aksu River Basin Oasis".

 

- Line 45 = Please elaborate and provide specific examples from around the world;

- Line 78 = I suggest deleting the last sentence and combining the remaining paragraph with the previous one, as they are similar in terms of assessing the validity of the manuscript;

- Line 83 = I propose to combine chapters 2 and 3 as "Materials and methods" to better organize the manuscript;

- Line 84 = Subsection 2.1 I propose to call "Study area and Environmental Setting";

- Line 104 = I propose to name subsection 2.2 "Sampling etc...". Additionally, please more elaborate on the sample collection process. How were the samples transported to the laboratory, etc.?;

- Line 121 = I suggest that Chapter 3 "Methods" be seamlessly combined with the previous one;

- Line 158, 172, 180 = The risk indicators cited should be better highlighted. Perhaps a subsections like "2.2.1; 2.2.2" or simply a new paragraph starting with the name of that indicator in the sentence;

- Line 218 = It would be necessary to improve the quality of "Figure 2" or split it into more figures [e.g. a), b), c), d)] to increase their resolution;

- Line 237 = Why mainly in the eastern oasis ?;

- Line 257 = The quality of "Figure 5" should also be improved as it is illegible and it is difficult to read the list;

- Line 262 = Please describe this reason in more detail, why, what factors caused such differences in concentrations?;

- Line 286, 328 = Move paragraph to "Conclusions";

- Line 336 = I suggest moving figure b) down and enlarging both figures a) and b) to improve their quality.

 

In summary, the manuscript requires further thorough editing and improvement to be suitable for final publication.

 

Author Response

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

 

Reviewer #1

 

Comments 1: Line 45 = Please elaborate and provide specific examples from around the world;

Response 1: Thank you for the valuable suggestion. We have added specific global examples of high-fluoride groundwater in the Introduction. These additions strengthen the global and regional context of the study. For details, please refer to the introduction section.

“An assessment incorporating over 400,000 fluoride measurements identified high-risk hotspots in central Australia, western North America, eastern Brazil, and extensive areas across Africa and Asia. Approximately 180 million people are potentially affected worldwide, with the majority residing in Asia and Africa. In China, high-fluoride groundwater is widely distributed in arid and semi-arid regions such as the Tarim Basin, Hetao Plain, and the North China Plain, where fluoride concentrations frequently exceed the WHO guideline of 1.5 mg/L [1].”

[1].Podgorski, J.; Berg, M. Global Analysis and Prediction of Fluoride in Groundwater. Nat. Commun. 2022, 13, 4232.

 

Comments 2: Line 78 = I suggest deleting the last sentence and combining the remaining paragraph with the previous one, as they are similar in terms of assessing the validity of the manuscript;

Response 2: We have fully adopted the expert's suggestion and have deleted the last sentence. We have also combined the remaining paragraphs with the previous ones.

 

Comments 3: Line 83 = I propose to combine chapters 2 and 3 as "Materials and methods" to better organize the manuscript;

Response 3: We have combined Chapter 2 and Chapter 3 into a single chapter titled "Materials and Methods" based on the expert's advice.

 

Comments 4: Line 84 = Subsection 2.1 I propose to call "Study area and Environmental Setting";

Response 4: We have renamed Section 2.1 as "Study area and Environmental Setting".

 

Comments 5: Line 104 = I propose to name subsection 2.2 "Sampling etc...". Additionally, please more elaborate on the sample collection process. How were the samples transported to the laboratory, etc.?;

Response 5: Following the experts' advice, we have changed the title of Section 2.2 to "Sampling and Laboratory Analysis", and have also added specific details regarding sample transportation and storage, including sealed packaging, refrigerated transportation, and laboratory storage conditions.

 

Comments 6: Line 121 = I suggest that Chapter 3 "Methods" be seamlessly combined with the previous one;

Response 6: We have adopted the experts' suggestion and combined Chapter 2 and Chapter 3 into "Materials and Methods", achieving a seamless transition.

 

Comments 7: Line 158, 172, 180 = The risk indicators cited should be better highlighted. Perhaps a subsections like "2.2.1; 2.2.2" or simply a new paragraph starting with the name of that indicator in the sentence;

Response 7: We followed the expert's advice and added three sub-chapters (2.4.1, 2.4.2, 2.4.3) to the health risk assessment section to clearly highlight each risk indicator. 

 

Comments 8: Line 218 = It would be necessary to improve the quality of "Figure 2" or split it into more figures [e.g. a), b), c), d)] to increase their resolution;

Response 8: We have fully adopted the expert's suggestions and split the original Figure 2 into four sub-figures. We have also enlarged the font size in the figure and increased the image resolution to ensure clear readability.

 

Comments 9: Line 237 = Why mainly in the eastern oasis ?;

Response 9: We have provided supplementary explanations in the results section, adding a brief explanation of the causes of the high fluoride areas in the eastern oases, which are attributed to low terrain, shallow groundwater level, and strong evaporation and concentration effects.

 

Comments 10: Line 257 = The quality of "Figure 5" should also be improved as it is illegible and it is difficult to read the list;

Response 10: We have taken the experts' advice and re-generated Figure 5. We have used vector format and enlarged the font size of the text in the figure to ensure that the legend and text are clearly legible.

 

Comments 11: Line 262 = Please describe this reason in more detail, why, what factors caused such differences in concentrations?;

Response 11: We have, based on the expert's opinions, provided detailed explanations in the result analysis for the reasons behind the spatial differences in fluoride concentrations in soil and groundwater. This is mainly attributed to the different geochemical environmental control factors for each. Please refer to the "3.1. Accuracy Evaluation and Spatial Distribution Patterns" section for more details.

 

Comments 12: Line 286, 328 = Move paragraph to "Conclusions";

Response 12: We have moved the paragraph to the "Conclusion" section as suggested by the expert.  

 

Comments 13: Line 336 = I suggest moving figure b) down and enlarging both figures a) and b) to improve their quality.

Response 13: The two panels of Figure 6 have been rearranged vertically and enlarged to improve clarity.

Reviewer 2 Report

Comments and Suggestions for Authors

Comments

The manuscript presents a large-scale, multi-media investigation of fluoride contamination in an arid oasis system, with robust sampling (2009 soil + 264 groundwater samples) and a multi-method analytical framework . The topic is highly relevant to the journal readership. However, several methodological, interpretative, and presentation issues must be addressed before publication

• The introduction lists objectives but no explicit hypothesis (e.g., "We hypothesize that fluoride enrichment is controlled by evapotranspiration rather than anthropogenic sources"
• The authors cite global and Chinese studies but omit recent work on Aksu River Basinitself
• RBF is used, but Kriging (universal or ordinary) is the geostatistical gold standard for environmental data. The authors compare RBF, LPI, and GPI but not Kriging, it is suggested to provide justification
• Table 2 is extremely dense and lacks citation columns. Some parameter values (e.g., PM10 = 0.1197 mg/m³) appear arbitrary.
• The ER index was originally designed for heavy metals using toxicity response factors. Fluoride is not a metal, and its "potential ecological risk" factor is not standard.So how the author is confident to apply the index for fluoride?
• Figure 1:Sample locations should be distinguished (soil vs. groundwater points). Also, add a scale bar and north arrow.
• Figure 3:The three panels (RBF, LPI, GPI) are hard to compare – use the same color scale for all.
• Figure 6 (heatmap):The correlation coefficients are hard to read; increase font size or use a matrix with values inside cells.
• Page 2, line 63: "groundwater- derived wastewater"which is unclear.

Comments on the Quality of English Language

its fine

Author Response

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

 

Reviewer #2

 

Comments 1: The introduction lists objectives but no explicit hypothesis (e.g., "We hypothesize that fluoride enrichment is controlled by evapotranspiration rather than anthropogenic sources".

Response 1: Thanks to the expert's advice, we have fully adopted it. We have added a clear scientific assumption at the end of the introduction: The fluorine enrichment in the oasis areas of the Akso River Basin is mainly controlled by natural hydrogeochemical processes (evaporation and alkaline activation), rather than direct human input.

 

Comments 2: The authors cite global and Chinese studies but omit recent work on Aksu River Basinitself.

Response 2: Thank you for the suggestion. We have added citations of recent studies specifically focused on the Aksu River Basin to better reflect the regional research context.

 

Comments 3: RBF is used, but Kriging (universal or ordinary) is the geostatistical gold standard for environmental data. The authors compare RBF, LPI, and GPI but not Kriging, it is suggested to provide justification.

Response 3: Thank you for the suggestion. A justification for not including Kriging has been added to the Methods section: the data deviate from normality and spatial autocorrelation structures are poorly fitted, and RBF performs better under such non-stationary and non-normal conditions. The comparison with LPI and GPI has been retained as they are computationally efficient and widely used in regional-scale mapping.

 

Comments 1: Table 2 is extremely dense and lacks citation columns. Some parameter values (e.g., PM10 = 0.1197 mg/m³) appear arbitrary.

Response 1: We have fully adopted the experts' suggestions. We have simplified Table 2, reduced the information density, and supplemented the source of the PM10 values (based on the annual average values of environmental monitoring in the Akso River Basin). At the same time, we have adjusted the numerical precision to two significant figures, making the table clearer and easier to read.

 

Comments 4: The ER index was originally designed for heavy metals using toxicity response factors. Fluoride is not a metal, and its "potential ecological risk" factor is not standard.So how the author is confident to apply the index for fluoride?

Response 4: Thank you for raising this important point. We have added a justification in the Methods section for applying the ER index to fluoride, citing precedents in the literature and clarifying that ER is used as a semi-quantitative indicator of potential ecological stress rather than an absolute measure of toxicity, with Tr=1 adopted from established studies.

 

Comments 5: Figure 1:Sample locations should be distinguished (soil vs. groundwater points). Also, add a scale bar and north arrow.

Response 5: Thank you for the suggestion. Figure 1 has been revised to distinguish soil and groundwater sampling points with different symbols, and a scale bar and north arrow have been added. The caption has been updated accordingly.

 

Comments 6: Figure 3:The three panels (RBF, LPI, GPI) are hard to compare – use the same color scale for all.

Response 6: Thank you for the suggestion. The color scales in Figure 3 have been unified across the three interpolation methods for soil and groundwater separately, allowing direct visual comparison.

 

Comments 7: Figure 6 (heatmap):The correlation coefficients are hard to read; increase font size or use a matrix with values inside cells.

Response 7: Thank you for the expert's advice. We have enlarged the font size of Figure 6, moved Figure b down, and improved the image resolution, which has significantly enhanced the readability.

 

Comments 8: Page 2, line 63: "groundwater- derived wastewater"which is unclear.

Response 8: Thank you for pointing out the unclear wording. “Groundwater-derived wastewater” has been replaced with “via groundwater-irrigated agriculture” to more accurately describe the exposure pathway.

Reviewer 3 Report

Comments and Suggestions for Authors

1.Section 3.2, Lines 180-195: The "Reference Dose Allocation Factor (SAF=0.5)" is introduced in the calculation formulas and parameter table, but no justification for its necessity or the basis for the value of 0.5 is provided. For a single pollutant risk assessment, introducing SAF requires sufficient rationale; otherwise, it may compromise the model's clarity and the comparability of the results.

2.Section 3.3, Lines 196-215: The analysis of driving mechanisms relies solely on Spearman's correlation analysis. Correlation does not imply causation. Relying only on this method makes it difficult to deeply "reveal the dominant enrichment mechanisms." The methodological section should objectively acknowledge this limitation or supplement it with methods such as machine learning models to quantify the relative importance of driving factors.

3.Section 4.2, Lines 273-285: The report only states that the mean Hazard Quotient (HQ) for children via the drinking water pathway is 1.83 (>1). It fails to provide spatial distribution information of the risk or the proportion exceeding the standard (e.g., the percentage of sampling points or areas where HQ > 1). This limits the value of the conclusions for guiding risk management and control.

4.Section 5.2, Lines 387-397: A reasonable explanation is provided here for the seemingly contradictory observation that soil fluoride is "unpolluted" (low Igeo) but poses a "moderate ecological risk" (high ER). However, this key concept—the fundamental difference in evaluation perspectives between Igeo and ER—was not clarified in advance in the Methods section (Section 3.2). This omission can easily confuse readers when they encounter the results.

5.Section 5.3, Lines 398-413: Listing "insufficient exploration of the impacts of human activities" as a limitation is inaccurate. Throughout the study's data collection and driver selection (Table 3), no indicators of human activities (e.g., land use type, irrigation intensity) were included. This constitutes a gap in the research design itself, not merely an issue of "insufficient exploration."

6.Line 444 (Data Availability Statement): The statement is listed as "Not applicable." However, the data supporting this study originates from a national geological survey project and constitutes key evidence for the paper's conclusions. This declaration does not align with the trends of open science or the data accessibility requirements of most journals.

7.Inconsistent Terminology and Formatting Throughout the Text: There are inconsistencies in terminology and minor formatting issues. For example, the English descriptions of the dominant mechanisms in the Abstract and Discussion sections slightly differ (e.g., "convergence–concentration–evaporation" vs. "sink concentration"). A thorough review of the entire text is needed to unify key terminology and ensure accurate labeling in figures and tables.

Author Response

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

 

Reviewer #3

 

Comments 1: Section 3.2, Lines 180-195: The "Reference Dose Allocation Factor (SAF=0.5)" is introduced in the calculation formulas and parameter table, but no justification for its necessity or the basis for the value of 0.5 is provided. For a single pollutant risk assessment, introducing SAF requires sufficient rationale; otherwise, it may compromise the model's clarity and the comparability of the results.

Response 1: We have fully adopted the expert's suggestions. We have added the basis for the value of SAF = 0.5, stating that it comes from the default assumption of the US EPA and is used to distribute the reference dose under various exposure pathways. Even for a single pollutant, retaining this parameter is to maintain consistency with the standard HRA protocol. For details, please refer to the "2.4.3. Human Health Risk Assessment" section.

 

Comments 2: Section 3.3, Lines 196-215: The analysis of driving mechanisms relies solely on Spearman's correlation analysis. Correlation does not imply causation. Relying only on this method makes it difficult to deeply "reveal the dominant enrichment mechanisms." The methodological section should objectively acknowledge this limitation or supplement it with methods such as machine learning models to quantify the relative importance of driving factors.

Response 2: We have, in accordance with the experts' suggestions, added an objective explanation of the limitations of the correlation analysis in the methodology section, acknowledging that it cannot infer causal relationships, and recommend the use of machine learning models for verification in the future. For details, please refer to the "2.5. Driving Factor Analysis" section.

 

Comments 3: Section 4.2, Lines 273-285: The report only states that the mean Hazard Quotient (HQ) for children via the drinking water pathway is 1.83 (>1). It fails to provide spatial distribution information of the risk or the proportion exceeding the standard (e.g., the percentage of sampling points or areas where HQ > 1). This limits the value of the conclusions for guiding risk management and control.

Response 3: We have fully adopted the experts' suggestions and have supplemented the proportion of sampling points where the HQ value for children through drinking water is greater than 1 (31.44%), as well as the spatial distribution characteristics of the risks. The high-risk areas are highly consistent with the hotspots of fluoride in groundwater.

 

Comments 4: Section 5.2, Lines 387-397: A reasonable explanation is provided here for the seemingly contradictory observation that soil fluoride is "unpolluted" (low Igeo) but poses a "moderate ecological risk" (high ER). However, this key concept—the fundamental difference in evaluation perspectives between Igeo and ER—was not clarified in advance in the Methods section (Section 3.2). This omission can easily confuse readers when they encounter the results.

Response 4: We have, based on the expert's suggestions, provided additional explanations in the methodology section regarding the differences in the assessment perspectives of Igeo and ER, in order to prevent confusion among readers in the results section. For details, please refer to the "2.4.2. Potential Ecological Risk Index Evaluation" section.

 

Comments 5: Section 5.3, Lines 398-413: Listing "insufficient exploration of the impacts of human activities" as a limitation is inaccurate. Throughout the study's data collection and driver selection (Table 3), no indicators of human activities (e.g., land use type, irrigation intensity) were included. This constitutes a gap in the research design itself, not merely an issue of "insufficient exploration."

Response 5: We have revised the relevant statements based on the experts' suggestions, clearly acknowledging that the omission of human activity indicators is a limitation of the research design itself, rather than a simple "insufficient exploration". Future studies should explicitly incorporate these variables. For details, please refer to the "4.3. Limitations and Future Research Directions" section.

 

Comments 6: Line 444 (Data Availability Statement): The statement is listed as "Not applicable." However, the data supporting this study originates from a national geological survey project and constitutes key evidence for the paper's conclusions. This declaration does not align with the trends of open science or the data accessibility requirements of most journals.

Response 6: We have fully adopted the expert's suggestion and modified the data availability statement to comply with the journal's requirements, which reads: "The raw data supporting the conclusions of this article will be made available by the authors on request."

 

Comments 7: Inconsistent Terminology and Formatting Throughout the Text: There are inconsistencies in terminology and minor formatting issues. For example, the English descriptions of the dominant mechanisms in the Abstract and Discussion sections slightly differ (e.g., "convergence–concentration–evaporation" vs. "sink concentration"). A thorough review of the entire text is needed to unify key terminology and ensure accurate labeling in figures and tables.

Response 7: We have fully adopted the expert's suggestions, conducted a thorough review throughout the document, and standardized the key terms. The mechanism of fluoride enrichment in groundwater is uniformly expressed as "convergence-concentration-evaporation mechanism". And ensure that the labels in the charts and tables are accurate and error-free.

Reviewer 4 Report

Comments and Suggestions for Authors

• The authors state, “In contrast, groundwater fluoride levels were generally low, with >90% of areas containing <2.0 mg·L⁻¹.” This is not adequately phrased, because the limit for fluoride in drinking water is 1.5 ppm, while the Chinese Class III groundwater standard is 1.0 ppm. Therefore, concentrations below 2 ppm cannot be described as generally low.
• The authors state, “Based on the ‘Geochemical Evaluation Specifications for Land Quality (DZ/T 0295-2016)’ and ‘Sanitary Standards for Drinking Water (GB 5749-2022),’ fluoride concentrations in soil and groundwater were classified into five levels.” However, these levels are not specified anywhere in the manuscript. Accordingly, the results should be discussed in line with the defined classification levels.
• In Table 5, it is not appropriate for the column heading to read “Sampling point.” A term such as “Concentration” should be used instead, because concentration is what is actually being presented.
• What value was used for C0 for fluoride in groundwater in the ER calculation? If 1 ppm was used, then the ER calculation is not correct. For groundwater, the maximum concentration is 8.9000 mg/L, yet ER is reported as 9.0816; the minimum is 0.1859 mg/L, yet ER is reported as 0.1897; and the average concentration is 1.6037 mg/L, yet ER is reported as 1.6364. Groundwater Igeo values also cannot be cleanly reproduced from the stated equation and k = 1.5. The calculations should therefore be checked carefully.
• When Figure 3 is examined, it is not clear what is actually being presented. It should be stated somewhere in the legend that fluoride concentration is being shown.
• Section 2.2 is entitled “Figures, Tables and Schemes,” although it actually describes data acquisition. Sections 3.2 and 3.3 are both entitled “Formatting of Mathematical Components,” although one section describes risk equations and the other driver analysis. These titles should be replaced with accurate scientific headings.
• An incorrect term, “geoconcentration index,” is used for Igeo, and a second incorrect term, “bioconcentration index,” is later used for the same concept.
• The interpolation-model selection is not reported adequately. R² is stated as a selection metric, but R² values are not actually presented in the text. Only RMSE values are given. The model-comparison subsection should therefore be expanded to provide all stated performance metrics, the validation design, and a clear rationale for the final model choice.
• The health-risk modelling choices are not justified sufficiently. The manuscript combines soil oral, dermal, and inhalation routes with groundwater drinking-water exposure, but it does not explain why food-chain uptake was discussed in the introduction and then omitted from the quantitative framework.
• Table 5 is considered unusable in its current form. The table heading indicates “Oral Ingestion,” but the listed routes are “Skin Contact,” “Respiratory Inhalation,” and “Water Ingestion,” whereas the groundwater block lacks a route name entirely. In addition, the total non-carcinogenic risk, that is, the hazard index, should be provided for soil. For groundwater, it would be identical to HQ ingestion, because only one exposure route was considered.
• Figure 6 is presented in such a way that no information is actually shown on the Spearman correlation coefficients between fluoride and the other parameters, because fluoride concentration does not appear in the matrix in Figure 6, even though this is discussed in the text with reference to Figure 6.
• Excessive numerical precision is used throughout, especially in Tables 4 and 5. Values such as 1.0098, 3.0205, 1.3987, 9.0816, 0.3409, and 0.8543 imply a level of precision that is not justified by environmental sampling variability and the limited analytical detail provided. Rounding to two or, at most, three significant figures is recommended.
• The section dealing with sampling and laboratory analysis is not described in sufficient detail. The sampling description is incomplete, instrumental specifications are missing, and QA/QC reporting is insufficient. Please include these information in the revised manuscript.
• What is missing from the manuscript, and what would strengthen it, is a comparison of fluoride concentrations with those reported for other locations in China and worldwide. This would substantially strengthen the discussion.

Author Response

I would like to express my heartfelt respect and gratitude to the expert for your meticulous and professional revision in your busy schedule, which makes the article more rigorous and further improved. Thanks to the editor and experts for the opportunity to revise. If there is any problem, please feel free to contact me at any time. I am very willing to make positive changes.

 

Reviewer #4

 

Comments 1: The authors state, “In contrast, groundwater fluoride levels were generally low, with >90% of areas containing <2.0 mg·L⁻¹.” This is not adequately phrased, because the limit for fluoride in drinking water is 1.5 ppm, while the Chinese Class III groundwater standard is 1.0 ppm. Therefore, concentrations below 2 ppm cannot be described as generally low.

Response 1: We have fully adopted the expert's suggestions and revised the expressions in the text. "In contrast, the fluoride content in groundwater is generally low. More than 68.56% of the areas have a fluoride content lower than 1.0 milligram per liter." For details, please refer to the "3.1. Accuracy Evaluation and Spatial Distribution Patterns" section.

 

Comments 2: The authors state, “Based on the ‘Geochemical Evaluation Specifications for Land Quality (DZ/T 0295-2016)’ and ‘Sanitary Standards for Drinking Water (GB 5749-2022),’ fluoride concentrations in soil and groundwater were classified into five levels.” However, these levels are not specified anywhere in the manuscript. Accordingly, the results should be discussed in line with the defined classification levels.

Response 2: We followed the expert's suggestions and specifically elaborated on the five-level classification standards for soil and groundwater fluoride in the manuscript. Additionally, in the result analysis, the relevant results were discussed according to the defined grading standards. For details, please refer to the "3.1. Accuracy Evaluation and Spatial Distribution Patterns" section.

 

Comments 3: In Table 5, it is not appropriate for the column heading to read “Sampling point.” A term such as “Concentration” should be used instead, because concentration is what is actually being presented.

Response 3: We have adopted the expert's suggestion and changed the title of the "Sampling point" column in Table 5 to "Concentration" to accurately reflect the content of the table.

 

Comments 4: What value was used for C0 for fluoride in groundwater in the ER calculation? If 1 ppm was used, then the ER calculation is not correct. For groundwater, the maximum concentration is 8.9000 mg/L, yet ER is reported as 9.0816; the minimum is 0.1859 mg/L, yet ER is reported as 0.1897; and the average concentration is 1.6037 mg/L, yet ER is reported as 1.6364. Groundwater Igeo values also cannot be cleanly reproduced from the stated equation and k = 1.5. The calculations should therefore be checked carefully.

Response 4: We have recalculated the ER and Igeo values of groundwater fluoride based on the experts' suggestions, and corrected the erroneous data in Table 4. The maximum ER value has been revised to 8.90, and the other values have also been adjusted accordingly. 

 

Comments 5: When Figure 3 is examined, it is not clear what is actually being presented. It should be stated somewhere in the legend that fluoride concentration is being shown.

Response 5: We have taken the experts' advice and added the information about fluoride concentration to the legend, and also modified the title of the picture.

 

Comments 6: Section 2.2 is entitled “Figures, Tables and Schemes,” although it actually describes data acquisition. Sections 3.2 and 3.3 are both entitled “Formatting of Mathematical Components,” although one section describes risk equations and the other driver analysis. These titles should be replaced with accurate scientific headings.

Response 6: We have fully adopted the experts' suggestions and revised the chapter titles to ensure they accurately reflect the content. “Figures, Tables and Schemes” has been changed to “Sampling and Laboratory Analysis”. “Formatting of Mathematical Components” has been changed to “Health and Ecological Risk Assessment” and “Driving Factor Analysis”.

 

Comments 7: An incorrect term, “geoconcentration index,” is used for Igeo, and a second incorrect term, “bioconcentration index,” is later used for the same concept.

Response 7: The incorrect terms “geoconcentration index” and “bioconcentration index” have been replaced throughout with the correct term “geo-accumulation index (Igeo)”.

 

Comments 8: The interpolation-model selection is not reported adequately. R² is stated as a selection metric, but R² values are not actually presented in the text. Only RMSE values are given. The model-comparison subsection should therefore be expanded to provide all stated performance metrics, the validation design, and a clear rationale for the final model choice.

Response 8: We followed the experts' suggestions and added the R² values of three methods, the cross-validation method (10-fold random sampling), and clarified the basis for choosing RBF. For details, please refer to the "3.1. Accuracy Evaluation and Spatial Distribution Patterns" section.

 

Comments 9: The health-risk modelling choices are not justified sufficiently. The manuscript combines soil oral, dermal, and inhalation routes with groundwater drinking-water exposure, but it does not explain why food-chain uptake was discussed in the introduction and then omitted from the quantitative framework.

Response 9: We incorporated the expert's opinions and provided supplementary explanations in the manuscript. The reasons for excluding the food chain route have been clarified: the lack of data on fluoride concentrations in crops and the bioaccumulation coefficients, and the need for specialized agricultural sampling, which is beyond the scope of this study. For details, please refer to "2.4.3. Human Health Risk Assessment".

 

Comments 10: Table 5 is considered unusable in its current form. The table heading indicates “Oral Ingestion,” but the listed routes are “Skin Contact,” “Respiratory Inhalation,” and “Water Ingestion,” whereas the groundwater block lacks a route name entirely. In addition, the total non-carcinogenic risk, that is, the hazard index, should be provided for soil. For groundwater, it would be identical to HQ ingestion, because only one exposure route was considered.

Response 10: We have redesigned Table 5 based on the experts' suggestions, adding the soil hazard index (HI) and clarifying each exposure route. For groundwater, the only exposure route is drinking water, and the HI is equal to HQ.

 

Comments 11: Figure 6 is presented in such a way that no information is actually shown on the Spearman correlation coefficients between fluoride and the other parameters, because fluoride concentration does not appear in the matrix in Figure 6, even though this is discussed in the text with reference to Figure 6.

Response 11: Figure 6 has been regenerated to include fluoride concentration as a variable in the correlation matrix, with correlation coefficients displayed.

 

Comments 12: Excessive numerical precision is used throughout, especially in Tables 4 and 5. Values such as 1.0098, 3.0205, 1.3987, 9.0816, 0.3409, and 0.8543 imply a level of precision that is not justified by environmental sampling variability and the limited analytical detail provided. Rounding to two or, at most, three significant figures is recommended.

Response 12: We have fully adopted the expert's suggestion and rounded the figures to two decimal places. 

 

Comments 13: The section dealing with sampling and laboratory analysis is not described in sufficient detail. The sampling description is incomplete, instrumental specifications are missing, and QA/QC reporting is insufficient. Please include these information in the revised manuscript.

Response 13: We have fully adopted the experts' suggestions and supplemented the QA/QC information, including standard substances, blank samples, parallel samples, RSD, recovery rates and instrument models. For details, please refer to the "2.2. Sampling and Laboratory Analysis" section.

 

Comments 14: What is missing from the manuscript, and what would strengthen it, is a comparison of fluoride concentrations with those reported for other locations in China and worldwide. This would substantially strengthen the discussion.

Response 14: We have adopted the experts' suggestions and added comparisons of fluoride concentrations with those in other high-fluoride areas both at home and abroad in the result analysis, including the East African Rift Valley, the Hetao Plain, the Tarim Basin, the Yuncheng Basin, and Guizhou, etc. For details, please refer to the "3.1. Accuracy Evaluation and Spatial Distribution Patterns" section.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors The authors responded to the recommendations.

Author Response

Thank you very much!

Reviewer 2 Report

Comments and Suggestions for Authors

the paper can be accepted now

Comments on the Quality of English Language

its fine

Author Response

Thank you very much!

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have responded well to my questions and concerns and have significantly improved the quality of the manuscript. However, there are still certain aspects that have not been adequately addressed. Once these issues are resolved, the manuscript will be ready for publication. My comments are below.

• Although it is stated in response that fluoride concentration was included in the Spearman correlation matrix, it is still not present there. Instead, fluoride can only be discussed through Mantel’s p and Mantel’s r. From the figure, it is not possible to determine the actual correlation coefficient. Additionally, in the figure, the color bar is labeled as Pearson’s r, while the authors state that they used Spearman correlation. It therefore remains unclear whether Pearson or Spearman correlation was actually applied.
• The authors have chosen to present results rounded to two decimal places, regardless of the appropriate number of significant figures. As a result, the risk values for dermal contact and inhalation are reported as 0.00, which is not acceptable. This does not mean that the risk is zero, but rather that an inappropriate level of numerical precision has been applied.

Author Response

Comments 1: Although it is stated in response that fluoride concentration was included in the Spearman correlation matrix, it is still not present there. Instead, fluoride can only be discussed through Mantel’s p and Mantel’s r. From the figure, it is not possible to determine the actual correlation coefficient. Additionally, in the figure, the color bar is labeled as Pearson’s r, while the authors state that they used Spearman correlation. It therefore remains unclear whether Pearson or Spearman correlation was actually applied.

Response 1: Thank you for the correction. Figure 6 has been regenerated to clearly display Spearman’s correlation coefficients (ρ) between fluoride concentrations and all driving factors, with values shown inside each cell. The color bar label has been corrected from “Pearson’s r” to “Spearman’s ρ”, and the caption has been updated accordingly. The Methods section has also been verified to specify Spearman’s rank correlation.

 

Comments 2: The authors have chosen to present results rounded to two decimal places, regardless of the appropriate number of significant figures. As a result, the risk values for dermal contact and inhalation are reported as 0.00, which is not acceptable. This does not mean that the risk is zero, but rather that an inappropriate level of numerical precision has been applied.

Response 2: Thank you for pointing out this critical issue. Previously rounding to two decimal places resulted in dermal contact and inhalation risk values being displayed as 0.00, which incorrectly implies zero risk. We have now restored the values to three or four significant figures and added a note to the table explaining the precision principle. These small but non-zero risk values are now appropriately represented.