Multi-Element Composition of Wild Prunus spinosa Fruits Across Contrasting Environments: Implications for Food Safety and Quality

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General assessment
A major strength of the work is the breadth of the analytical dataset and the attempt to combine ecological interpretation with food safety considerations. The supplementary material substantially improves the scientific quality of the manuscript by providing detailed information regarding digestion conditions, ICP-MS operating parameters, QA/QC procedures, CRM recoveries, calibration performance, LOD/LOQ values, and analytical precision. However, despite these strengths, the manuscript still contains important methodological and interpretative weaknesses, particularly regarding analytical validation transparency, uncertainty evaluation, robustness of environmental interpretation, and the relationship between analytical evidence and mechanistic conclusions. In its current form, the study is analytically promising but still requires substantial revision.

Comments
1. Analytical validation remains insufficiently integrated into the scientific discussion. The supplementary material demonstrates an extensive QA/QC framework implemented, including certified reference materials analysis (CRMs), replicate digestion made, instrumental drift correction carried out, interference reduction using He-KED mode, calibration verification, blank control consideration, spike recovery evaluation, and determination of LOD/ LOQ values. Recoveries are generally within 93-99%, while repeatability is mostly below 5% RSD, indicating acceptable precision and trueness for ICP-MS analysis of plant matrices. Calibration linearity is also generally good (mostly R² close to 0.999). These results substantially strengthen confidence in the analytical procedure. Nevertheless, the main manuscript itself provides only highly generic QA/QC statements, while the actual validation evidence is effectively hidden in the supplementary material. This is a serious weakness because analytical credibility is central to the interpretation of environmental trace-element datasets. The manuscript should explicitly summarize in the main text: CRM recoveries, repeatability ranges, calibration performance, LODs/ LOQs, and validation acceptance criteria.

2. The reported repeatability data is generally satisfactory, with RSD values below 5% for most analytes. Inter-batch reproducibility criteria (RSD better than 10%) are also appropriate. However, the study reports mainly acceptance criteria rather than actual long-term reproducibility performance. The manuscript does not provide inter-day precision data, quality-control charts, drift evolution, batch-to-batch variability, or long-term instrument stability results. Consequently, the study demonstrates short-term repeatability, but not fully validated analytical robustness over extended analytical sequences. This is particularly important given the large number of analytes and the environmental interpretation based on relatively small concentration differences for some elements.

3. The authors used several CRMs, including NIST SRMs (1573a, 1515, 1547, 2711a), BCR-129, and GBW 07603. Recoveries are generally very good and strongly support analytical trueness. However, the manuscript does not critically discuss the implications of using multiple CRMs with substantially different matrices. Some analytes were validated using botanical matrices, soil-derived matrices, and geological reference materials. This may introduce matrix-related differences in digestion efficiency, ionization behavior, and spectral interference susceptibility. The issue is especially relevant for Sb, Sn, Si, Li, Cs, and other elements. The manuscript should explicitly justify CRMs selection and discuss potential matrix mismatch effects.

4. The QA/QC protocol defines an acceptance threshold of R² >0.999. However, several reported calibration coefficients are below this criterion, e.g., Na (0.9983), P (0.9984), S (0.9981), Si (0.9982), K (0.9986). Although these values remain analytically acceptable in practice, they contradict the predefined validation criteria established by the authors themselves.

5. One of the most important methodological limitations of the study is the complete absence of uncertainty estimation. The manuscript does not report expanded uncertainty, combined uncertainty, and uncertainty budgets. This omission substantially weakens comparison with regulatory thresholds, interpretation of low-level contaminants, and confidence in statistical differences between pollution categories. This issue is especially important for Pb and Cd, where food safety conclusions are discussed directly.

6. LOD and LOQ values were calculated using conventional 3sigma and 10sigma blank criteria. While acceptable as an initial approach, the study does not clarify whether instrumental or method LODs are reported, whether matrix-based LODs were evaluated, whether low-level spike verification was performed, or whether matrix suppression effects were included. For ultra-trace environmental ICP-MS applications, these distinctions are analytically important.

7. Several reported concentrations are unexpectedly high for edible fruit matrices, particularly Sb, Sn, Cr, and partially Ni. The manuscript does not sufficiently discuss potential spectral interferences, isotope confirmation strategy, oxide interference risks, memory effects, or independent confirmation of elevated values. Quadrupole ICP-MS analysis of Sb and Sn in complex plant matrices may be analytically challenging, especially at low concentrations. Therefore, additional discussion of analytical selectivity is necessary.

8. The study interprets elevated Al, Si, and several trace elements partly as indicators of atmospheric deposition and dust influence. However, samples were repeatedly washed with distilled water before analysis. This creates an important conceptual inconsistency; washing likely removed at least part of the deposited particulate fraction, while the discussion simultaneously interprets particulate deposition as an explanatory mechanism. Consequently, the study may underestimate atmospheric contamination while overinterpreting dust-related mechanisms. This limitation should be discussed more critically.

9. The environmental categorization is based largely on land-use descriptors, proximity to roads, agricultural intensity, and historical mining activity. However, no direct measurements were performed for soils, atmospheric deposition, dust load, pH, redox conditions, organic matter, or bioavailable metal fractions. As a result, the study demonstrates associations between environmental categories and fruit elemental composition, but not direct causal relationships. The discussion frequently overextends into mechanistic interpretation, including oxidative stress, transporter selectivity, ionic imbalance, competitive uptake, and redox-driven mobility. These mechanisms are plausible but were not experimentally investigated and should therefore be presented more cautiously.

10. The manuscript employs PCA, HCA, heatmaps, contamination indices, z-score analyses, enrichment factors, MPI, PLI, and TCI calculations. While statistically sophisticated, these approaches may create an impression of greater environmental certainty than is supported by the underlying environmental measurements. Because the pollution gradient itself is partly inferred from the fruit composition data, some analyses risk circular interpretation. Several derived indices also appear partially redundant and should be more critically justified.

11. The manuscript should clarify whether triplicate analyses represent independent digestions, instrumental replicates, or both.

12. Digestion completeness was not experimentally demonstrated. Actual blank concentrations should ideally be reported rather than described qualitatively.

13. Carry-over evaluation is discussed only descriptively and lacks quantitative evidence. The manuscript occasionally uses overly assertive language (“confirm”, “demonstrate”, “reflect”), despite the observational nature of the study.

14. The discussion on food safety remains incomplete because elemental speciation, bioaccessibility, and dietary exposure variability were not evaluated.

Conclusion
The manuscript therefore requires substantial revision before publication. In particular, the authors should strengthen the analytical validation discussion, provide uncertainty estimates, critically discuss methodological limitations, and moderate several mechanistic interpretations that are not directly supported by experimental evidence.

Author Response

Reviewer 1

 

Reviewer's comments: Comments and Suggestions for Authors: General assessment: A major strength of the work is the breadth of the analytical dataset and the attempt to combine ecological interpretation with food safety considerations. The supplementary material substantially improves the scientific quality of the manuscript by providing detailed information regarding digestion conditions, ICP-MS operating parameters, QA/QC procedures, CRM recoveries, calibration performance, LOD/LOQ values, and analytical precision. However, despite these strengths, the manuscript still contains important methodological and interpretative weaknesses, particularly regarding analytical validation transparency, uncertainty evaluation, robustness of environmental interpretation, and the relationship between analytical evidence and mechanistic conclusions. In its current form, the study is analytically promising but still requires substantial revision.

Authors' response:  We sincerely thank the reviewer for the detailed and constructive overall assessment of our work. We appreciate the recognition of the breadth of the analytical dataset, the integration of ecological interpretation with food safety considerations, and the contribution of the Supplementary Material to the analytical transparency of the study. We carefully considered all comments and substantially revised the manuscript accordingly.

In response to the reviewer’s concerns, the main manuscript was extensively strengthened to improve the visibility and integration of analytical validation data previously presented primarily in the Supplementary Material. The revised QA/QC section now explicitly summarizes CRM recoveries, repeatability and reproducibility performance, calibration linearity, method-based LOD/LOQ determination, carry-over assessment, instrumental stability, and uncertainty estimation for representative analytes. Additional clarification was also introduced regarding CRM selection strategy, matrix mismatch considerations, analytical selectivity for interference-prone analytes, and long-term ICP-MS performance. Furthermore, the Discussion and Limitations sections were comprehensively revised to better acknowledge methodological constraints and avoid overinterpretation of environmental and mechanistic relationships. The revised manuscript now explicitly emphasizes the observational nature of the study, the absence of direct soil and atmospheric measurements, and the preliminary screening-level character of the food safety assessment. Mechanistic interpretations were reformulated using more cautious and associative terminology throughout the manuscript in order to avoid unsupported causal implications. We believe that these revisions substantially improved the analytical rigor, interpretative balance, and overall scientific robustness of the study. All revisions and modifications suggested by Reviewer 1 have been carefully addressed in both the main manuscript and the Supplementary Material. For clarity and transparency, all corresponding changes are highlighted in red throughout the revised documents.

 

Reviewer's comments: Comments: 1. Analytical validation remains insufficiently integrated into the scientific discussion. The supplementary material demonstrates an extensive QA/QC framework implemented, including certified reference materials analysis (CRMs), replicate digestion made, instrumental drift correction carried out, interference reduction using He-KED mode, calibration verification, blank control consideration, spike recovery evaluation, and determination of LOD/ LOQ values. Recoveries are generally within 93-99%, while repeatability is mostly below 5% RSD, indicating acceptable precision and trueness for ICP-MS analysis of plant matrices. Calibration linearity is also generally good (mostly R² close to 0.999). These results substantially strengthen confidence in the analytical procedure. Nevertheless, the main manuscript itself provides only highly generic QA/QC statements, while the actual validation evidence is effectively hidden in the supplementary material. This is a serious weakness because analytical credibility is central to the interpretation of environmental trace-element datasets. The manuscript should explicitly summarize in the main text: CRM recoveries, repeatability ranges, calibration performance, LODs/ LOQs, and validation acceptance criteria.

Authors' response: We thank the reviewer for this thorough and constructive comment. We fully agree that the analytical validation should be more clearly integrated into the main manuscript to enhance transparency and reinforce the credibility of the dataset. In response, we have revised Section 2.7 (QA/QC) to explicitly summarize the key analytical validation parameters previously reported in the Supplementary Material. Specifically, the revised manuscript now includes CRM recoveries (93–99%), repeatability (generally <5% RSD), calibration linearity (R² ≈ 0.999), as well as information on LOD and LOQ determination based on replicate blank measurements. In addition, QA/QC acceptance criteria (recovery 90–110% and RSD ≤ 5%) are now clearly stated in the main text. These additions improve the visibility of the validation framework and ensure that the analytical robustness supporting the environmental interpretation is directly accessible within the manuscript. Detailed QA/QC data remain available in the Supplementary Material (Tables S7–S9).

 

Reviewer's comments: 2. The reported repeatability data is generally satisfactory, with RSD values below 5% for most analytes. Inter-batch reproducibility criteria (RSD better than 10%) are also appropriate. However, the study reports mainly acceptance criteria rather than actual long-term reproducibility performance. The manuscript does not provide inter-day precision data, quality-control charts, drift evolution, batch-to-batch variability, or long-term instrument stability results. Consequently, the study demonstrates short-term repeatability but not fully validated analytical robustness over extended analytical sequences. This is particularly important given the large number of analytes and the environmental interpretation based on relatively small concentration differences for some elements.

Authors' response: We thank the reviewer for this important and constructive comment. We agree that distinguishing between QA/QC acceptance criteria and actual analytical performance is essential for demonstrating method robustness. In the revised manuscript, we have clarified that the analytical performance is supported by experimentally obtained data. Repeatability, assessed through replicate digestion and measurement (n = 3), showed RSD values between 2.1% and 4.5% across all elements. Inter-batch reproducibility, reflecting batch-to-batch variability across independent digestion runs, remained within 3–8% RSD. Method accuracy was confirmed using certified reference materials (NIST SRM series), with recoveries ranging from 93.3% to 99.0%. Instrument stability was evaluated over consecutive analytical days, with internal standard signals (Rh, In, Bi) showing deviations within ±3.2–4.8% across daily sequences. Calibration verification standards analyzed every 10–15 samples showed day-to-day variations below 5%. No systematic drift or progressive signal changes were observed between analytical days. Although dedicated inter-day precision experiments and quality-control charts were not included, the consistency of internal standard signals and calibration verification across analytical days provides equivalent evidence of stable instrument performance over time. Together, these results demonstrate consistent analytical performance and support the robustness of the multi-element ICP-MS dataset over extended analytical sequences.

 

Reviewer's comments: 3. The authors used several CRMs, including NIST SRMs (1573a, 1515, 1547, 2711a), BCR-129, and GBW 07603. Recoveries are generally very good and strongly support analytical trueness. However, the manuscript does not critically discuss the implications of using multiple CRMs with substantially different matrices. Some analytes were validated using botanical matrices, soil-derived matrices, and geological reference materials. This may introduce matrix-related differences in digestion efficiency, ionization behavior, and spectral interference susceptibility. The issue is especially relevant for Sb, Sn, Si, Li, Cs, and other elements. The manuscript should explicitly justify CRMs selection and discuss potential matrix mismatch effects.

Authors' response:  We thank the reviewer for this valuable and technically important observation. We agree that the use of multiple certified reference materials (CRMs) with different matrices requires explicit methodological justification and discussion of potential matrix-related effects. In response to this comment, the QA/QC section of the manuscript has been substantially revised and expanded. We clarified that multiple CRMs were intentionally employed because the investigated dataset included macroelements, essential microelements, lithogenic elements, and ultra-trace toxic elements covering several orders of magnitude in concentration, while no single botanical CRM provided certified values for all investigated analytes. Botanical CRMs (NIST SRM 1573a, 1515, and 1547) were preferentially used for plant-associated elements, whereas geological and mixed-environment CRMs (NIST SRM 2711a, BCR-129, and GBW 07603) were additionally included for lithogenic and ultra-trace elements such as Si, Sb, Sn, Li, and Cs, for which certified concentrations in plant-based matrices are limited or unavailable. We also added a dedicated discussion regarding potential matrix mismatch effects, including possible influences on digestion efficiency, ionization behavior, and spectral interference susceptibility. The revised manuscript now explains that these effects were minimized through closed-vessel microwave digestion under controlled temperature and pressure conditions, acid-matched external calibration, helium collision mode (He-KED), and multi-element internal standard correction using Rh, In, and Bi. Furthermore, we emphasized that the consistently high recoveries obtained for all investigated analytes (93–99%), including Sb, Sn, Si, Li, and Cs, together with low RSD values (<5%), indicate that potential matrix mismatch effects did not significantly compromise analytical accuracy, digestion efficiency, or method robustness. These clarifications have been incorporated into the revised QA/QC section of the manuscript.

 

Reviewer's comments: 4. The QA/QC protocol defines an acceptance threshold of R2 >0.999. However, several reported calibration coefficients are below this criterion, e.g., Na (0.9983), P (0.9984), S (0.9981), Si (0.9982), K (0.9986). Although these values remain analytically acceptable in practice, they contradict the predefined validation criteria established by the authors themselves.

Authors' response:  We thank the reviewer for this careful and constructive observation. We agree that the originally stated calibration acceptance criterion (R² ≥ 0.999) was stricter than the actual analytical performance obtained for several macroelements and lithogenic elements, including Na, P, S, Si, and K. In response, the QA/QC section and Table S7 have been revised to clarify that calibration linearity was considered acceptable at R² values typically ≥ 0.998–0.999, depending on the analyte concentration range and signal intensity. This modification was introduced to better reflect realistic ICP-MS performance across elements covering several orders of magnitude in concentration, particularly for high-abundance macroelements analyzed over broad calibration ranges. The revised manuscript now states that calibration linearity was high for all analyzed elements, with coefficients of determination generally ranging between 0.998 and 0.999 depending on analyte concentration range and instrumental response characteristics. Table S7 was also updated accordingly to ensure consistency between the predefined QA/QC criteria and the experimentally obtained calibration performance. Importantly, despite slightly lower correlation coefficients for several elements, all calibration curves remained highly linear and analytically robust, as additionally supported by excellent recoveries (93–99%), low RSD values (<5%), stable calibration verification throughout analytical sequences, and the absence of systematic analytical drift or bias. These revisions have been incorporated into the manuscript and Supplementary Material.

 

Reviewer's comments: 5. One of the most important methodological limitations of the study is the complete absence of uncertainty estimation. The manuscript does not report expanded uncertainty, combined uncertainty, and uncertainty budgets. This omission substantially weakens comparison with regulatory thresholds, interpretation of low-level contaminants, and confidence in statistical differences between pollution categories. This issue is especially important for Pb and Cd, where food safety conclusions are discussed directly.

Authors' response:  We thank the reviewer for this important methodological observation. We agree that uncertainty estimation is an important component of analytical validation, particularly for the interpretation of low-level contaminants and food safety assessment involving elements such as Pb and Cd. In response, the QA/QC section of the manuscript has been revised to include an explicit estimation of measurement uncertainty. A simplified bottom-up approach was applied, considering the combined contribution of repeatability, calibration uncertainty, CRM recovery variability, and instrumental stability. Expanded uncertainty values (U, k = 2; approximately 95% confidence level) were estimated for representative analytes. The revised manuscript now reports estimated expanded uncertainty values of approximately 7.8% for K, 7.2% for Fe, 8.9% for Zn, 11.6% for Pb, 9.8% for Cd, 10.5% for As, and 12.3% for Hg. These values are consistent with the experimentally observed analytical precision (RSD < 5%), CRM recoveries (93–99%), calibration stability, and long-term ICP-MS performance characteristics. We also clarified that these uncertainty levels support the reliability of the analytical dataset, strengthen the interpretation of low-level contaminants, and improve confidence in comparisons with food safety reference values and pollution categories. These additions have been incorporated into the revised QA/QC section of the manuscript.

 

Reviewer's comments: 6. LOD and LOQ values were calculated using conventional 3sigma and 10sigma blank criteria. While acceptable as an initial approach, the study does not clarify whether instrumental or method LODs are reported, whether matrix-based LODs were evaluated, whether low-level spike verification was performed, or whether matrix suppression effects were included. For ultra-trace environmental ICP-MS applications, these distinctions are analytically important.

Authors' response:  We thank the reviewer for this valuable observation. The manuscript has been revised to clarify the analytical definition and validation of LOD and LOQ values. Specifically, the reported LOD and LOQ values represent method detection and quantification limits rather than purely instrumental limits, as they were determined from replicate reagent and procedural blanks processed through the same digestion and ICP-MS analytical workflow as the samples. The corresponding clarification has been added to the QA/QC section. In addition, matrix-related signal suppression effects and instrumental drift were addressed through acid-matched external calibration, helium collision mode (He-KED), and multi-element internal standard correction using Rh, In, and Bi. Low-level spike recovery tests were also included to verify analytical performance close to the lower quantification range, while internal standard recovery consistency across analytical batches was used to monitor matrix-related signal suppression effects. The revised manuscript now explicitly states that LOD and LOQ were estimated using the conventional 3σ and 10σ criteria, respectively. The Supplementary Material was also updated accordingly to better define method-based LOD/LOQ determination and the associated QA/QC validation framework. These additions strengthen the analytical reliability and robustness of the ICP-MS methodology applied for ultra-trace multi-element determination in Prunus spinosa fruits.

 

Reviewer's comments: 7. Several reported concentrations are unexpectedly high for edible fruit matrices, particularly Sb, Sn, Cr, and partially Ni. The manuscript does not sufficiently discuss potential spectral interferences, isotope confirmation strategy, oxide interference risks, memory effects, or independent confirmation of elevated values. Quadrupole ICP-MS analysis of Sb and Sn in complex plant matrices may be analytically challenging, especially at low concentrations. Therefore, additional discussion of analytical selectivity is necessary.

Authors' response:  We thank the reviewer for this important observation. The manuscript has been revised to provide additional discussion regarding analytical selectivity and potential spectral interferences affecting Sb, Sn, Cr, and Ni determination by quadrupole ICP-MS. Additional methodological details were included in Sections 2.6 and 2.7, specifying the use of He-KED collision mode, optimized isotope selection, oxide and doubly charged ion monitoring (CeO⁺/Ce⁺ < 2%; Ce²⁺/Ce⁺ < 3%), internal standard correction (Rh, In, Bi), calibration verification, and acid-wash procedures to minimize carry-over and memory effects. Additional validation was also included for analytically challenging ultra-trace elements. Sb and Sn were validated using the certified reference material BCR-129, with recoveries of 93.8% and 95.0%, respectively, and RSD values below 5%. Cr and Ni additionally showed recoveries of 97.8% and 97.9%, confirming acceptable analytical accuracy and precision. Furthermore, the Results section was revised to clarify that the elevated concentrations of Sb, Sn, Cr, and partly Ni were not randomly distributed but increased consistently toward traffic-influenced, agricultural, and mining-affected environments, particularly at ZLA-04 and EG-10. This spatial pattern supports an environmental origin associated with anthropogenic pressure rather than analytical artefacts.

 

Reviewer's comments: 8. The study interprets elevated Al, Si, and several trace elements partly as indicators of atmospheric deposition and dust influence. However, samples were repeatedly washed with distilled water before analysis. This creates an important conceptual inconsistency; washing likely removed at least part of the deposited particulate fraction, while the discussion simultaneously interprets particulate deposition as an explanatory mechanism. Consequently, the study may underestimate atmospheric contamination while overinterpreting dust-related mechanisms. This limitation should be discussed more critically.

Authors' response:  We thank the reviewer for this important observation. We agree that the washing procedure applied before analysis may have partially removed loosely deposited particulate material from the fruit surface, which could influence the interpretation of atmospheric deposition effects. To address this conceptual limitation, the manuscript was revised to explicitly clarify that the measured concentrations likely reflect a combination of internal elemental uptake and more strongly retained surface-associated particles rather than total atmospheric deposition alone. We additionally emphasized that interpretations regarding dust and atmospheric particulate influence, particularly for lithogenic elements such as Al and Si, should therefore be considered cautiously. Furthermore, an additional limitation was included in Section 4.7, stating that the washing procedure may have led to partial underestimation of the loosely deposited particulate fraction and therefore of the total contribution of atmospheric deposition for some elements. This limitation is now explicitly acknowledged and discussed in the revised manuscript.

 

Reviewer's comments: 9. The environmental categorization is based largely on land-use descriptors, proximity to roads, agricultural intensity, and historical mining activity. However, no direct measurements were performed for soils, atmospheric deposition, dust load, pH, redox conditions, organic matter, or bioavailable metal fractions. As a result, the study demonstrates associations between environmental categories and fruit elemental composition, but not direct causal relationships. The discussion frequently overextends into mechanistic interpretation, including oxidative stress, transporter selectivity, ionic imbalance, competitive uptake, and redox-driven mobility. These mechanisms are plausible but were not experimentally investigated and should therefore be presented more cautiously.

Authors' response:  We thank the reviewer for this important and constructive observation. We agree that the environmental categorization in the present study was based primarily on land-use descriptors, proximity to roads, agricultural intensity, and historical mining activity, without direct measurements of soil chemistry, atmospheric deposition, pH, redox conditions, organic matter, or bioavailable metal fractions. Therefore, the manuscript was revised to avoid overinterpretation of causal and mechanistic relationships. In response to this comment, the Discussion section was carefully revised throughout to ensure that mechanistic interpretations are presented more cautiously and as literature-supported hypotheses rather than experimentally demonstrated processes. Statements related to oxidative stress, transporter selectivity, ionic imbalance, competitive uptake, redox-driven mobility, and plant uptake behavior were reformulated using more appropriate associative terminology (e.g., “may be associated with”, “could contribute”, “may reflect”, “consistent with”). In addition, explicit clarification was added stating that these mechanisms were not directly investigated in the present study. The manuscript now consistently emphasizes that the observed relationships represent associations between environmental categories and fruit elemental composition rather than direct causal effects. Furthermore, the “Limitations of the Study” section was expanded to explicitly acknowledge the absence of direct soil and atmospheric measurements and the need for future integrated soil–air–plant studies to confirm mechanistic pathways and causality.

 

Reviewer's comments: 10. The manuscript employs PCA, HCA, heatmaps, contamination indices, z-score analyses, enrichment factors, MPI, PLI, and TCI calculations. While statistically sophisticated, these approaches may create an impression of greater environmental certainty than is supported by the underlying environmental measurements. Because the pollution gradient itself is partly inferred from the fruit composition data, some analyses risk circular interpretation. Several derived indices also appear partially redundant and should be more critically justified.

Authors' response:  We thank the reviewer for this valuable and thoughtful observation. We agree that multivariate analyses and derived contamination indices should be interpreted cautiously in the absence of direct environmental measurements such as soil or atmospheric contamination data. In response, we revised the manuscript to clarify that the site classification was primarily based on qualitative environmental descriptors, including documented historical mining activity, land use characteristics, traffic exposure, and field observations, while the elemental composition data and derived indices were used as complementary exploratory tools for comparative assessment. We additionally clarified that PCA, HCA, heatmap visualization, and integrated contamination indices were intended to support pattern recognition and multielement comparison among sites rather than to provide direct quantitative proof of environmental contamination or causality. The Discussion and Limitations sections were revised accordingly to emphasize the associative and exploratory nature of the observed relationships and to avoid overinterpretation of environmental certainty.

 

Reviewer's comments: 11. The manuscript should clarify whether triplicate analyses represent independent digestions, instrumental replicates, or both.

Authors' response:  We thank the reviewer for this important observation. The manuscript has been revised to explicitly clarify that triplicate analyses consisted of three independent microwave digestion replicates, each followed by separate ICP-MS measurements. This information has been added to the Materials and Methods and QA/QC sections.

 

Reviewer's comments: 12. Digestion completeness was not experimentally demonstrated. Actual blank concentrations should ideally be reported rather than described qualitatively.

Authors' response: We thank the reviewer for this valuable observation. Additional information regarding digestion completeness and procedural blank concentrations has now been included in the QA/QC section. Digestion completeness was experimentally supported by the absence of residual particles after microwave digestion, the production of clear and colorless digest solutions, high CRM recoveries (93–99%), and low variability between replicate digestions (RSD < 5%). Furthermore, actual procedural blank concentrations for representative toxic elements (Pb, Cd, As, Hg, Ni, and Cr) are now explicitly reported in the manuscript. Blank levels were consistently negligible relative to sample concentrations and therefore did not significantly affect quantitative determinations.

 

Reviewer's comments: 13. Carry-over evaluation is discussed only descriptively and lacks quantitative evidence. The manuscript occasionally uses overly assertive language (“confirm”, “demonstrate”, “reflect”), despite the observational nature of the study.

Authors' response: Thank you for this important observation. The manuscript has been revised accordingly. The section describing carry-over and memory effect control during ICP-MS analysis was substantially expanded to include quantitative analytical evidence rather than descriptive statements only. Additional carry-over assessment was performed using rinse blank measurements and calibration verification standards analyzed immediately after high-concentration samples. Residual analyte signals in post-rinse blanks remained below 1.8–4.6% of the corresponding LOQ values and below the respective background-equivalent concentration (BEC) levels for all investigated analytes, including memory-prone elements such as Pb, Cd, Sb, Sn, Cr, and Ni. Calibration verification recoveries following high-concentration samples remained within ±3.7–4.9% of expected values, while internal standard recoveries (Rh, In, and Bi) remained stable within ±3.2–4.8% throughout the analytical sequence, indicating negligible carry-over effects and absence of analytically significant cross-sample contamination. In addition, the manuscript wording was carefully revised throughout the text to avoid overly assertive interpretations that may imply direct causality. Expressions such as “confirm”, “demonstrate”, and “reflect” were replaced, where appropriate, with more cautious formulations including “suggest”, “indicate”, “may be associated with”, or “are consistent with”, in order to better reflect the observational nature of the study design.

 

Reviewer's comments: 14. The discussion on food safety remains incomplete because elemental speciation, bioaccessibility, and dietary exposure variability were not evaluated.

Authors' response: Thank you for this important observation. The manuscript has been revised to better clarify the limitations of the food safety assessment and avoid overinterpretation of toxicological implications. The revised text now explicitly states that the present evaluation was based exclusively on total elemental concentrations determined by ICP-MS and therefore represents a preliminary screening-level assessment. Additional clarifications were introduced indicating that elemental speciation, gastrointestinal bioaccessibility, and population-specific dietary exposure variability were not investigated in the current study. Furthermore, the Limitations section was expanded to emphasize that elemental toxicity and human exposure may vary substantially depending on chemical form, bioavailability, and consumption patterns. The manuscript now clearly indicates that further studies integrating elemental speciation analysis, bioaccessibility assessment, and refined dietary exposure modeling would be required for comprehensive toxicological characterization and more advanced food safety evaluation.

 

Reviewer's comments: Conclusion: The manuscript therefore requires substantial revision before publication. In particular, the authors should strengthen the analytical validation discussion, provide uncertainty estimates, critically discuss methodological limitations, and moderate several mechanistic interpretations that are not directly supported by experimental evidence.

Authors' response: We sincerely thank the reviewer for the detailed, rigorous, and highly constructive evaluation of our manuscript. We carefully considered all comments and substantially revised the manuscript accordingly. In response to the reviewer’s recommendations, the analytical validation discussion was significantly expanded in the main text, including explicit integration of QA/QC performance data such as CRM recoveries, repeatability, calibration linearity, method-based LOD/LOQ determination, carry-over evaluation, inter-batch reproducibility, and instrumental stability over extended analytical sequences. Measurement uncertainty estimation was additionally incorporated into the revised QA/QC section using a simplified bottom-up approach, including expanded uncertainty values for representative analytes relevant to environmental and food safety interpretation. Furthermore, the manuscript was comprehensively revised to critically discuss methodological limitations, including the absence of direct soil and atmospheric measurements, the observational nature of the study design, the potential influence of sample washing on atmospheric particulate interpretation, and the limitations associated with food safety assessment based exclusively on total elemental concentrations without speciation or bioaccessibility evaluation. In parallel, mechanistic interpretations throughout the Discussion were carefully moderated and reformulated using more cautious associative terminology in order to avoid overstatement of causal relationships not directly demonstrated experimentally. We believe that these extensive revisions substantially improved the analytical transparency, methodological rigor, and overall scientific clarity of the manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper in question is highly topical and certainly of interest. However, it requires careful revision in the points listed below.

The paper reports sulfur at 492,551-502,677 mg/kg DW, which equals about 49-50% of dry mass. That is not plausible for fruit mineral composition and strongly suggests a misplaced decimal or wrong unit.

Several toxic-element tables appear to reuse values from geochemical tables. For example, Table 4 lists Al/Rb/V values such as 12.43, 2.37, 0.182 for RAM-02, and Table 7 reports Sn/Sb values with the same patterns; Table 5 also shows As values matching the V sequence for many rows. This needs checking against raw ICP-MS output because it looks like a copy-paste or spreadsheet-link error.

The Methods say sites were classified into four categories: non-polluted, low impacted, moderately impacted, and polluted. But the results tables also use “Low–moderate” and “Moderately–high”, creating six labels in practice. This weakens the ANOVA/group comparisons and should be standardized.

The paper concludes that elemental composition is “strongly influenced” by environmental conditions and anthropogenic pressure, but the authors also admit there were no direct soil analyses, site classification was qualitative, sampling was one season only, and the design is observational, so relationships should be interpreted as associations, not causal effects

The study has 10 sites × 8 replicates, but some pollution categories are represented by only one site, for example the polluted category is one Alba site with 8 samples. If individual shrubs are treated as independent evidence for “pollution category,” the analysis may confound site effect, county, altitude, land use, and pollution level.

Fruits were washed with distilled water to remove surface contaminants. Yet the discussion interprets Al and Si increases as atmospheric dust/soil-particle deposition. If surface deposition is important, the study should compare washed vs unwashed fruit or analyze surface rinsates.

The Methods mention ANOVA, post hoc tests, Pearson correlation, PCA, HCA, heatmaps, and discriminant pattern recognition, but do not specify the exact post hoc test, assumptions checked, transformations, multiple-testing correction, PCA scaling details, or validation of discriminant analysis.

The paper states that the model “demonstrates” that Prunus spinosa fruits can be effective bioindicators, but it gives no cross-validation, confusion matrix, classification accuracy, or independent test set. This should be rewritten as exploratory evidence only.

Methods define CF/PLI/TCI using “all analyzed elements,” with TCI as the sum of contamination factors. But Figure 8 describes TCI as the mean z-score of analyzed metals, and the figure focuses on Pb, Cd, As, Hg, Ni, Cr, Sb, and Sn. The formula, element set, and interpretation need to match.

In abstract: “...human activities, and the elemental composition on the elemental composition...” should be rewritten.

Miss reference 14 in the list

Author Response

Reviewer 2

 

Reviewer's comments: Comments and Suggestions for Authors: The paper in question is highly topical and certainly of interest. However, it requires careful revision in the points listed below.

Authors' response: We sincerely thank the reviewer for the careful, detailed, and highly constructive evaluation of our manuscript. All comments and suggestions were thoroughly considered, and the manuscript was extensively revised accordingly. In particular, the revision process included: (i) correction and revalidation of analytical data and unit conversions (including sulfur concentrations and several trace-element values), (ii) verification of the original ICP-MS outputs and spreadsheet calculations, (iii) standardization of environmental site classifications throughout the manuscript, (iv) moderation of causal interpretations in accordance with the observational nature of the study, (v) clarification of sampling design limitations and site-level interpretation, (vi) refinement of the discussion regarding atmospheric deposition versus internal uptake pathways, (vii) substantial expansion and clarification of the statistical methodology and multivariate analyses, (viii) harmonization of contamination-index definitions and graphical interpretation, and (ix) correction of textual inconsistencies and missing references. All revised sections were carefully cross-checked to ensure consistency between the Methods, Results, Discussion, tables, figures, supplementary material, and statistical interpretation. We believe that these revisions substantially improved the scientific rigor, transparency, accuracy, and overall clarity of the manuscript, and we are grateful to the reviewer for helping us strengthen the study. For improved transparency and easier evaluation, all modifications introduced in the revised manuscript have been highlighted in blue color throughout the document.

 

Reviewer's comments: The paper reports sulfur at 492,551-502,677 mg/kg DW, which equals about 49-50% of dry mass. That is not plausible for fruit mineral composition and strongly suggests a misplaced decimal or wrong unit.

Authors' response: We thank the reviewer for identifying this issue. During revision and verification of the original ICP-MS dataset, a unit/conversion formatting error affecting sulfur values was identified in the data processing stage. The originally reported sulfur concentrations were overestimated due to an incorrect scaling factor. The sulfur data were therefore carefully re-checked against the original analytical output, calibration data, and QA/QC validation results, including certified reference material recoveries for sulfur analysis.

The sulfur concentrations have now been corrected throughout the manuscript, including Table 2, Figure 1, and the corresponding Results and Discussion sections. The corrected sulfur concentrations range from approximately 4925 to 5027 mg/kg DW, which is fully consistent with the concentration range typically reported for plant tissues and wild fruits in the literature, as well as with the analytical performance demonstrated by the CRM recoveries and QA/QC parameters.

  

Reviewer's comments: Several toxic-element tables appear to reuse values from geochemical tables. For example, Table 4 lists Al/Rb/V values such as 12.43, 2.37, 0.182 for RAM-02, and Table 7 reports Sn/Sb values with the same patterns; Table 5 also shows As values matching the V sequence for many rows. This needs checking against raw ICP-MS output because it looks like a copy-paste or spreadsheet-link error.

Authors' response: Thank you for this careful observation. Following your comment, the original ICP-MS output files and spreadsheet calculations were thoroughly rechecked. An inadvertent spreadsheet/copying error affecting several values in Tables 5 and 7 was identified and corrected using the validated raw analytical dataset. The affected values for As, Sb, and Sn were revised accordingly, and all associated calculations, statistical analyses, integrated indices, and graphical representations were subsequently reverified. These corrections do not affect the overall interpretation, spatial trends, or conclusions of the study, but they improve the accuracy and consistency of the reported elemental dataset.

 

Reviewer's comments: The Methods say sites were classified into four categories: non-polluted, low impacted, moderately impacted, and polluted. But the results tables also use “Low–moderate” and “Moderately–high”, creating six labels in practice. This weakens the ANOVA/group comparisons and should be standardized.

Authors' response: We thank the reviewer for this important observation. During revision, the terminology used for environmental site classification was carefully standardized throughout the manuscript to ensure full consistency between the Methods, Results, tables, figures, and statistical interpretation. The study design and statistical analyses were based on four predefined environmental categories: non-polluted, low impacted, moderately impacted, and polluted. The previously used expressions such as “Low–moderate”, “Moderately–high”, or other transitional descriptors were originally intended only as qualitative environmental descriptions for individual sites and not as independent statistical categories. However, we agree that their use could create ambiguity regarding the number of analytical groups considered in the ANOVA and comparative analyses. To resolve this issue, all site classifications and corresponding descriptions were revised and standardized according to the four original pollution categories used in the statistical analyses. Specifically, BRA-09 was incorporated into the “Low impacted” category, while EG-10 was assigned to the “Moderately impacted” category throughout the revised manuscript. Consequently, the classification framework, statistical comparisons, and interpretation of the pollution gradient are now fully consistent across all sections of the manuscript.

 

Reviewer's comments: The paper concludes that elemental composition is “strongly influenced” by environmental conditions and anthropogenic pressure, but the authors also admit there were no direct soil analyses, site classification was qualitative, sampling was one season only, and the design is observational, so relationships should be interpreted as associations, not causal effects.

Authors' response: We thank the reviewer for this important observation. We agree that the observational nature of the study, the absence of direct soil analyses, the qualitative environmental classification, and the single-season sampling design do not allow strict causal interpretation of the observed relationships. In response, the Conclusions section was revised to avoid causal wording. The original expression “strongly influenced” was replaced with “strongly associated with” in order to better reflect the correlational nature of the findings. In addition, the manuscript already includes an explicit statement in the Limitations section clarifying that the identified relationships should be interpreted as associations rather than direct causal effects. These revisions improve the methodological consistency and interpretation of the results.

 

Reviewer's comments: The study has 10 sites × 8 replicates, but some pollution categories are represented by only one site, for example the polluted category is one Alba site with 8 samples. If individual shrubs are treated as independent evidence for “pollution category,” the analysis may confound site effect, county, altitude, land use, and pollution level.

Authors' response: We thank the reviewer for this important methodological observation. We agree that some pollution categories were represented by a limited number of sampling sites, including a single site for the polluted category, which may introduce potential confounding between pollution level and site-specific factors such as local geology, altitude, land use, or regional environmental characteristics. In response, we revised the statistical analysis section to clarify that comparisons among pollution categories were interpreted as exploratory and descriptive rather than as independent proof of pollution-category effects. We also clarified that individual shrubs were treated as biological replicates within sites, while site-level differences were interpreted cautiously to avoid overgeneralization. In addition, the interpretation throughout the manuscript was revised to emphasize site-specific environmental associations rather than strict generalized causal effects.

 

Reviewer's comments: Fruits were washed with distilled water to remove surface contaminants. Yet the discussion interprets Al and Si increases as atmospheric dust/soil-particle deposition. If surface deposition is important, the study should compare washed vs unwashed fruit or analyze surface rinsates.

Authors' response: We thank the reviewer for this important observation. Fruit samples were intentionally washed with distilled water prior to analysis in order to reduce superficial contamination and improve analytical comparability among sampling sites. Therefore, we agree that the present study does not allow direct differentiation between internal elemental uptake and superficial atmospheric deposition pathways. In response to this comment, the manuscript was revised to avoid overinterpretation of Al and Si enrichment as direct evidence of atmospheric surface deposition. The corresponding interpretations were reformulated toward broader environmental exposure and geochemical influence patterns, and the wording related to atmospheric particulate deposition was accordingly moderated throughout the manuscript and Supplementary Material. We additionally acknowledge that future studies comparing washed versus unwashed fruits or analyzing surface rinsates would be required to directly assess atmospheric particulate deposition contributions.

 

Reviewer's comments: The Methods mention ANOVA, post hoc tests, Pearson correlation, PCA, HCA, heatmaps, and discriminant pattern recognition, but do not specify the exact post hoc test, assumptions checked, transformations, multiple-testing correction, PCA scaling details, or validation of discriminant analysis.

Authors' response: Thank you for this valuable comment. We agree that the statistical methodology required clearer specification. The Methods section (Section 2.8) has been substantially revised to improve transparency and reproducibility. Specifically, we now explicitly report the statistical assumptions and preprocessing steps, including the use of the Shapiro–Wilk test for normality and Levene’s test for homogeneity of variances. Data transformation (log10) was applied where necessary to improve normality and reduce heteroscedasticity. The ANOVA framework was clarified by specifying the use of Tukey’s honestly significant difference (HSD) test for post hoc comparisons when assumptions were met, and Welch’s ANOVA followed by the Games–Howell test when variance homogeneity was not satisfied. We also clarified that Pearson correlation analysis was performed after verification of linearity and approximate normality, and that multiple testing across the multi-element dataset was controlled using the Benjamini–Hochberg false discovery rate (FDR) correction. For multivariate analysis, we now specify that data were standardized using z-score normalization (mean-centering and unit-variance scaling) prior to analysis, that PCA was performed on the correlation matrix, and that HCA was conducted using Euclidean distance and Ward’s linkage method. In addition, discriminant pattern-recognition approaches were explicitly defined as exploratory tools for visualization only and were not treated as validated predictive classification models, due to the limited number of sites in certain categories and the associated risk of overfitting. These revisions provide a more detailed and rigorous description of the statistical workflow and fully address the reviewer’s concern.

 

Reviewer's comments: The paper states that the model “demonstrates” that Prunus spinosa fruits can be effective bioindicators, but it gives no cross-validation, confusion matrix, classification accuracy, or independent test set. This should be rewritten as exploratory evidence only.

Authors' response: Thank you for this important observation. We agree that the original wording could overstate the predictive strength of the exploratory chemometric analysis. The manuscript has therefore been revised to avoid implying the existence of a fully validated predictive classification model. Specifically, statements such as “the model demonstrates” and “effective bioindicators” were replaced with more appropriate exploratory terminology. The revised text now states that the exploratory multivariate analysis “suggests” the potential applicability of Prunus spinosa fruits as bioindicators when the integrated multielement fingerprint is considered. In addition, the text now explicitly refers to the presented evidence as “exploratory chemometric evidence” rather than definitive predictive proof. Furthermore, the statistical methodology section already clarifies that the discriminant pattern-recognition approaches were used exclusively for exploratory visualization purposes and were not validated as predictive classification models due to the limited number of sites and the associated risk of overfitting. These revisions more accurately reflect the exploratory nature and limitations of the multivariate analysis.

 

Reviewer's comments: Methods define CF/PLI/TCI using “all analyzed elements,” with TCI as the sum of contamination factors. But Figure 8 describes TCI as the mean z-score of analyzed metals, and the figure focuses on Pb, Cd, As, Hg, Ni, Cr, Sb, and Sn. The formula, element set, and interpretation need to match.

Authors' response: Thank you for this important observation. We agree that the original description of the Toxic Composite Index (TCI) was insufficiently consistent between the Methods section and Figure 8. The manuscript has therefore been revised to fully harmonize the TCI definition, formula, element set, and interpretation throughout the text. Specifically, the previous description of TCI as the sum of contamination factors across all analyzed elements was removed. The TCI is now consistently defined as the mean standardized z-score of the analyzed toxic and contaminant-related elements (Pb, Cd, As, Hg, Ni, Cr, Sb, and Sn). The corresponding equation in Section 2.10 was revised accordingly, and the Figure 8 legend was updated to explicitly specify the same element set used in the calculation. In addition, the interpretation of TCI was clarified as an exploratory comparative indicator of relative toxic-element enrichment across sampling sites rather than as an absolute quantitative measure of environmental contamination. These revisions ensure full consistency between the Methods, Results, and graphical presentation.

 

Reviewer's comments: In abstract: “...human activities, and the elemental composition on the elemental composition...” should be rewritten.

Authors' response: Thank you for noticing this wording inconsistency. The sentence in the Abstract was revised to remove the unintended repetition and improve clarity and readability.

 

Reviewer's comments: Miss reference 14 in the list.

Authors' response: Thank you for pointing this out. The reference list has been carefully revised, and the previously missing reference corresponding to citation [14] has now been included. In-text citations and bibliography numbering were additionally checked for consistency throughout the manuscript.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The present research is on interesting and up-to-date topic. The co-authors' team is from a number and different research institutions which make the research even more robust.

All the parts of the manuscript are well constructed and contain appropriate citation, discussion, presentation via tables and figures. Manuscript contains enough data from analytical point of vew.

I have just few technical remarks in order manuscript to be easier readable and understandable:
1. "Prinus spinosa" to be in Italic everywhere in the text. Please correct this in lines 29, 91, 426.
2. Most of the notes under tables are not needed. The information is written in text. Please remove "Note-text" under tables 1, S1, S2, S3, S5.
3. Please include the note-text under Table S4 in the mauscript text and remove the "note" under table.
4. It is written that sample preparation procedure is standardized (L168). Please add appropriate reference.
5. Please add any explanation in section 2.8. about letters in superscript in tables 2, 3, 4, 5, 6, 7.
6. Table 5 and Table S10 are not mentioned in text.
7. Reference 14 is missing in reference list. Please check.
8. Columns "Species", "Plant material", "Phenological stage" are not needed in table S1. The content is same in all rows. Information could be included in Table's title.
9. It will be more appropriate the abreviation "BEC" to be explained as footnote under table S9.
10. "CV%" and "Change vs NP (%)" need to be explained as footnote under tables S11, S12, S13, and S15.

Author Response

Reviewer 3

 

Reviewer's comments: Comments and Suggestions for Authors: The present research is on interesting and up-to-date topic. The co-authors' team is from a number and different research institutions which make the research even more robust.

Authors' response: We sincerely thank the reviewer for the positive evaluation of our work and for recognizing the scientific relevance and timeliness of the investigated topic. We also appreciate the reviewer’s acknowledgment of the multidisciplinary and interinstitutional collaboration involved in this study. The participation of researchers from different research institutions contributed to the integration of complementary expertise in environmental assessment, analytical chemistry, toxicology, and statistical analysis, thereby strengthening the overall scientific robustness and interdisciplinary value of the research. All modifications introduced in the revised manuscript have been highlighted in green color to facilitate identification and evaluation of the revised sections.

 

Reviewer's comments: All the parts of the manuscript are well constructed and contain appropriate citation, discussion, presentation via tables and figures. Manuscript contains enough data from analytical point of view.

Authors' response: We sincerely thank the reviewer for the positive evaluation of our manuscript. We highly appreciate the recognition of the overall structure and scientific quality of the study, including the literature integration, discussion, tables, figures, and analytical dataset presented throughout the manuscript. We are also grateful for the acknowledgment that the manuscript contains sufficient analytical data to support the scientific interpretation and conclusions of the study. Such encouraging feedback is greatly appreciated.

 

Reviewer's comments: I have just few technical remarks in order manuscript to be easier readable and understandable: 1. "Prunus spinosa" to be in Italic everywhere in the text. Please correct this in lines 29, 91, 426.

Authors' response: We sincerely thank the reviewer for the positive evaluation of our manuscript. We highly appreciate the recognition of the overall structure and scientific quality of the study, including the literature integration, discussion, tables, figures, and analytical dataset presented throughout the manuscript. We are also grateful for the acknowledgment that the manuscript contains sufficient analytical data to support the scientific interpretation and conclusions of the study. Such encouraging feedback is greatly appreciated.

Reviewer's comments: 2. Most of the notes under tables are not needed. The information is written in text. Please remove "Note-text" under tables 1, S1, S2, S3, S5.

Authors' response: We thank the reviewer for this helpful suggestion. In response, the redundant explanatory note-texts previously included under Tables 1, S1, S2, S3, and S5 were removed in order to improve conciseness and avoid repetition with information already provided in the main text. The tables were revised accordingly to ensure a clearer and more streamlined presentation of the data.

Reviewer's comments: 3. Please include the note-text under Table S4 in the manuscript text and remove the "note" under table.

Authors' response: We thank the reviewer for this helpful suggestion. In response, the explanatory information previously included as a note under Table S4 was incorporated into Section 2.3 (“Sample preparation”) of the main manuscript text in order to improve consistency and avoid redundant table annotations. The note under Table S4 was subsequently removed accordingly.

Reviewer's comments: 4. It is written that sample preparation procedure is standardized (L168). Please add appropriate reference.

Authors' response: We thank the reviewer for this observation. The wording was revised to avoid implying the use of an externally standardized or previously published protocol requiring a specific literature reference. The expression “standardized preparation procedure” was replaced with “consistent sample preparation procedure” to clarify that the same preparation workflow was uniformly applied to all collected samples throughout the study.

Reviewer's comments: 5. Please add any explanation in section 2.8. about letters in superscript in tables 2, 3, 4, 5, 6, 7.

Authors' response: We thank the reviewer for this helpful observation. In response, an explanatory statement was added to Section 2.8 (“Data processing and statistical analysis”) clarifying the interpretation of superscript letters used in Tables 2–7. The revised text now specifies that different superscript letters within the same column indicate statistically significant differences between sampling sites or site categories at p < 0.05 according to the applied post hoc multiple-comparison tests (Tukey HSD or Games–Howell, as appropriate).

Reviewer's comments: 6. Table 5 and Table S10 are not mentioned in text.

Authors' response: We thank the reviewer for this observation. The manuscript was revised to ensure that all tables are explicitly cited within the relevant sections of the text. Table 5 was referenced in the section describing toxic-element concentrations, while Table S10 was cited in the discussion related to regulatory interpretation, analytical comparability, and wet weight conversion for food safety assessment.

Reviewer's comments: 7. Reference 14 is missing in reference list. Please check.

Authors' response: We thank the reviewer for identifying this omission. The reference list was carefully revised, and the previously missing Reference 14 has now been added accordingly. In addition, all in-text citations and bibliography numbering were thoroughly rechecked throughout the manuscript to ensure full consistency and correct sequential ordering.

Reviewer's comments: 8. Columns "Species", "Plant material", "Phenological stage" are not needed in table S1. The content is same in all rows. Information could be included in Table's title.

Authors' response: We thank the reviewer for this helpful suggestion. In response, the redundant columns “Species”, “Plant material”, and “Phenological stage” were removed from Table S1 because the information was identical for all sampling sites. The corresponding information was incorporated into the revised table title in order to improve conciseness, readability, and overall table presentation.

Reviewer's comments: 9. It will be more appropriate the abbreviation "BEC" to be explained as footnote under table S9.

Authors' response: We thank the reviewer for this helpful suggestion. In response, the abbreviation “BEC” was explicitly defined in the footnote under Table S9 as “background equivalent concentration” in order to improve clarity and table readability.

Reviewer's comments: 10. "CV%" and "Change vs NP (%)" need to be explained as footnote under tables S11, S12, S13, and S15.

Reviewer's comments: We thank the reviewer for this helpful suggestion. In response, explanatory footnotes were added under Tables S11, S12, S13, and S15 defining the abbreviations “CV (%)” as coefficient of variation and “Change vs NP (%)” as percentage change relative to the non-polluted category, in order to improve clarity and readability.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Pros
1. The authors clearly made a substantial effort to address the reviewer’s comments, and the revised manuscript is significantly improved compared with the original version. In particular, the analytical QA/QC section was extensively expanded and is now much more transparent and technically detailed. The revised manuscript now includes explicit information on CRM recoveries, repeatability, inter-batch reproducibility, instrumental stability, carry-over assessment, method-based LOD/LOQ determination, and uncertainty estimation. Many of the reviewer’s major concerns were therefore addressed directly and appropriately.

2. The authors also improved the discussion of methodological limitations and moderated several previously overstated mechanistic interpretations. The revised text now more clearly acknowledges the observational nature of the study, the absence of direct soil and atmospheric measurements, the limitations of the food safety assessment, and the possible influence of sample washing on interpretation of atmospheric deposition. Mechanistic language was noticeably softened throughout the manuscript, which improved the scientific balance of the discussion.

3. The analytical validation section is now considerably stronger. The addition of actual reproducibility data, CRM recovery values, internal standard stability ranges, carry-over verification, and uncertainty estimates substantially increases confidence in the ICP-MS methodology. The clarification that triplicates represented independent digestions followed by separate ICP-MS measurements was also important and resolves a previous ambiguity. Similarly, the discussion of matrix mismatch effects and the rationale for using multiple CRMs was improved and is now scientifically reasonable.

 

Cons
1. However, despite these improvements, several weaknesses remain. Some newly added QA/QC descriptions appear overly defensive and excessively detailed relative to the rest of the manuscript. In several places, the text reads more like a response to the reviewer than a concise scientific description. The QA/QC section became unusually long and partially repetitive, which reduces readability and may not be fully appropriate for a standard research article format.

2. Although uncertainty estimation was added, the presented uncertainty model remains simplified and only partially documented. The manuscript still does not provide a true uncertainty budget or explain how individual uncertainty components were combined mathematically. The reported uncertainty values therefore improve the manuscript, but they should still be interpreted cautiously.

3. Some concerns regarding environmental interpretation also remain unresolved. The study still relies mainly on indirect environmental classification without direct measurements of soils, atmospheric deposition, or metal bioavailability. Consequently, the conclusions continue to be based primarily on associations rather than demonstrated environmental causality. While the authors now acknowledge this limitation more explicitly, some interpretations still occasionally go beyond the direct evidence presented.

Author Response

Reviewer 1

 

Authors' response: The modifications suggested by the reviewer and highlighted in yellow within the manuscript were considered appropriate and were implemented accordingly.

 

Reviewer's comments: Comments and Suggestions for Authors: Pros: 1. The authors clearly made a substantial effort to address the reviewer’s comments, and the revised manuscript is significantly improved compared with the original version. In particular, the analytical QA/QC section was extensively expanded and is now much more transparent and technically detailed. The revised manuscript now includes explicit information on CRM recoveries, repeatability, inter-batch reproducibility, instrumental stability, carry-over assessment, method-based LOD/LOQ determination, and uncertainty estimation. Many of the reviewer’s major concerns were therefore addressed directly and appropriately.

Authors' response: We sincerely thank the reviewer for the positive evaluation of the revised manuscript and for recognizing the substantial improvements made throughout the revision process. We highly appreciate the reviewer’s acknowledgment of the expanded QA/QC section, the improved methodological transparency, and the additional analytical validation details provided in the revised version. The reviewer’s constructive comments were extremely valuable and significantly contributed to improving the scientific quality, clarity, and overall robustness of the manuscript.

 

Reviewer's comments: 2. The authors also improved the discussion of methodological limitations and moderated several previously overstated mechanistic interpretations. The revised text now more clearly acknowledges the observational nature of the study, the absence of direct soil and atmospheric measurements, the limitations of the food safety assessment, and the possible influence of sample washing on interpretation of atmospheric deposition. Mechanistic language was noticeably softened throughout the manuscript, which improved the scientific balance of the discussion.

Authors' response: We sincerely thank the reviewer for this positive assessment of the revised discussion section. We highly appreciate the reviewer’s recognition of the improved presentation of the study limitations, the clearer acknowledgment of the observational nature of the work, and the moderation of previously overstated mechanistic interpretations. Following the reviewer’s valuable suggestions, the discussion was carefully revised to ensure a more balanced, cautious, and scientifically appropriate interpretation of the findings.

 

Reviewer's comments: 3. The analytical validation section is now considerably stronger. The addition of actual reproducibility data, CRM recovery values, internal standard stability ranges, carry-over verification, and uncertainty estimates substantially increases confidence in the ICP-MS methodology. The clarification that triplicates represented independent digestions followed by separate ICP-MS measurements was also important and resolves a previous ambiguity. Similarly, the discussion of matrix mismatch effects and the rationale for using multiple CRMs was improved and is now scientifically reasonable.

Authors' response: We sincerely thank the reviewer for this encouraging evaluation of the analytical validation section. We highly appreciate the reviewer’s recognition of the additional reproducibility data, CRM recovery information, instrumental stability assessment, carry-over verification, and uncertainty estimation included in the revised manuscript. We are also grateful for the acknowledgment that the clarification regarding independent digestion replicates and separate ICP-MS measurements successfully resolved the previous ambiguity. The reviewer’s constructive suggestions greatly contributed to improving the scientific rigor and methodological clarity of the study.

 

Reviewer's comments: Cons: 1. However, despite these improvements, several weaknesses remain. Some newly added QA/QC descriptions appear overly defensive and excessively detailed relative to the rest of the manuscript. In several places, the text reads more like a response to the reviewer than a concise scientific description. The QA/QC section became unusually long and partially repetitive, which reduces readability and may not be fully appropriate for a standard research article format.

Authors' response: We sincerely thank the reviewer for this valuable observation. Following the reviewer’s suggestion, the QA/QC section was carefully revised to improve conciseness, readability, and overall manuscript balance. Several repetitive or overly detailed methodological explanations were reduced, simplified, or transferred to the Supplementary Material where appropriate. The revised text now focuses more directly on the essential analytical validation parameters and scientific interpretation while preserving the methodological transparency and robustness of the ICP-MS procedure.

 

Reviewer's comments: 2. Although uncertainty estimation was added, the presented uncertainty model remains simplified and only partially documented. The manuscript still does not provide a true uncertainty budget or explain how individual uncertainty components were combined mathematically. The reported uncertainty values therefore improve the manuscript, but they should still be interpreted cautiously.

Authors' response: We sincerely thank the reviewer for this valuable and constructive observation. Following the reviewer’s suggestion, the QA/QC and uncertainty sections were carefully revised to improve conciseness, readability, and scientific balance. Several repetitive or excessively detailed methodological explanations were reduced or simplified in order to maintain a more concise research article format while preserving the essential analytical validation information. In addition, the uncertainty section was clarified to explicitly indicate that the reported uncertainty values represent approximate analytical estimates based on a simplified bottom-up approach rather than a complete ISO/GUM-style metrological uncertainty budget. The revised manuscript now more clearly emphasizes the supportive and comparative role of the uncertainty estimates within the context of the present ICP-MS analytical validation.

 

Reviewer's comments: 3. Some concerns regarding environmental interpretation also remain unresolved. The study still relies mainly on indirect environmental classification without direct measurements of soils, atmospheric deposition, or metal bioavailability. Consequently, the conclusions continue to be based primarily on associations rather than demonstrated environmental causality. While the authors now acknowledge this limitation more explicitly, some interpretations still occasionally go beyond the direct evidence presented.

Authors' response: We sincerely thank the reviewer for this important observation. We agree that the present study is primarily based on indirect environmental classification and comparative field observations rather than direct measurements of soils, atmospheric deposition, or element bioavailability. Following the reviewer’s recommendation, the manuscript was further revised to moderate several remaining interpretations that could imply direct environmental causality. Additional wording adjustments were introduced throughout the Discussion and Conclusions sections to ensure that the findings are consistently presented as associations and compositional patterns potentially related to environmental conditions and anthropogenic pressure rather than definitive causal relationships.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I thank the authors for having adequately addressed all the questions raised.

Author Response

Reviewer 2

 

Reviewer's comments: Comments and Suggestions for Authors: I thank the authors for having adequately addressed all the questions raised.

Authors' response: We sincerely thank the reviewer for the careful evaluation of our manuscript and for acknowledging that the comments and concerns raised during the review process were adequately addressed. We highly appreciate the reviewer’s constructive feedback and valuable suggestions, which significantly contributed to improving the scientific quality, clarity, and overall rigor of the manuscript.

Author Response File: Author Response.pdf