Spatiotemporal Dynamics and Driving Forces of Vegetation Net Primary Productivity on Hainan Island (2001–2022)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Vegetation Net Primary Production is an essential component of the carbon cycle in terrestrial ecosystems. However, many regional features are still not fully understood. Therefore, research on Vegetation Net Primary Production is relevant.

The peer-reviewed paper undoubtedly has a scientific novelty. The scientific novelty lies in the identification of temporal and spatial patterns in the evolution of Vegetation Net Primary Production and its determining factors. This was one of the research objectives that the authors successfully completed.

The authors concentrated their research on Hainan Island. This is a good choice for conducting research, as this region is insufficiently studied.

The introduction is written quite well. The authors justified the relevance and explained the current state of the problem. The research objectives are formulated in detail, clearly and clearly.

The research methodology is generally consistent with the stated objectives of the work. The authors use a wide range of spatial and temporal analysis methods, including trend analysis (Theil-Sen, Mann–Kendall), forecasting (Hurst index), factor analysis and their interactions (Geodetector), and structural modeling (PLS-SEM). This multi-level approach is methodologically sound for studying the dynamics of Vegetation Net Primary Production and its drivers at the regional level. At the same time, individual stages require a clearer justification, structuring, and clarification of the logic of integrating the methods used.

Strengths of the methodology:

1. Comprehensive research. The strongest side of the work is a combination of time analysis (CV, Sen slope, Mann–Kendall), sustainability and future trends analysis (Hurst index), spatial factor analysis (Geodetector), and causal modeling (PLS-SEM). This allows us to consider Vegetation Net Primary Production not only as a dynamic indicator, but also as the result of a complex interaction of natural and anthropogenic factors.
2. Using a long-term time series (2001-2022). The 22-year analysis period is sufficient to identify stable trends and makes the use of nonparametric methods statistically justified.
3. Clear classification of impact factors. Dividing the drivers into climatic, topographic, and anthropogenic increases the interpretability of the results and facilitates the subsequent integration of the findings into environmental management practice.

Weaknesses and methodological limitations:

1. Despite the large number of methods used, the methodology is not completely clear: how exactly are the results of trend analysis, Geodetector, and PLS-SEM logically related; are the results of some methods used as inputs or hypotheses for others? This gives the impression of parallel rather than hierarchically linked analysis.
2. Insufficient elaboration of the PLS-SEM methodology.

Although the conceptual model has been described, there is no information on checking the reliability and validity of measurement models; model quality criteria. This reduces the strictness of the interpretation of cause-and-effect relationships.

3. Insufficient description of data sources and quality. The methodology lacks a detailed discussion of input data uncertainties, possible remote sensing errors, and differences between climate and socio-economic data sources.

Recommendations for improving the methodology:

Clearly build a methodological logic. It is recommended to explicitly show: the sequence of application of methods, the role of each method in testing specific hypotheses.

Strengthen the methodological block of PLS-SEM. Add: model quality indicators, stability and sensitivity tests.

Substantiate the scale of the analysis. Preferably: discuss the impact of spatial resolution.

To clarify the interpretation of Hurst analysis. It is recommended to formulate conclusions about future trends more carefully, and explicitly indicate the limitations of the extrapolation approach.

Add a section on data uncertainties and limitations. This will increase scientific transparency and trust in the results. The research results are described clearly and quite fully. At the same time, the authors use good visualization. The paper contains 7 visual figures of acceptable quality and 3 informative tables. The figures do not require improvement. However, for a more complete understanding of the design of the study and the results of the study, I would like to recommend adding a figure that would reflect the overall design of the study and all the methods used, indicating which tasks they are needed for. This will significantly strengthen the paper.

The conclusion is generally logically structured and directly correlates with the three stated goals of the study. The authors consistently summarize the results obtained using trend analysis, geographic detectors, and PLS-SEM, which indicates a high degree of consistency between tasks and conclusions.

Nevertheless, despite its methodological completeness, the conclusion does not fully reflect the applied purpose of the study related to the assessment of the effectiveness of environmental remediation measures and their contribution to carbon neutrality strategies, which were clearly outlined in the formulation of the goal.

Comparison of goals and conclusions

1. To study the spatiotemporal dynamics and future trends of NPP (2001-2022) using trend analysis. Fully implemented. The conclusion clearly describes the time trend; provides a quantitative assessment of the rate of increase in NPP; reflects spatial differences; includes forecast categories of future changes. This point is the strong point of the conclusion and correlates well with the task at hand.
2. Identify the main NPP drivers in terms of topography, climate, vegetation, and anthropogenic factors using Geodetector. Implemented as a whole adequately. The conclusion correctly highlights the dominant role of topography and climate; lists key variables (elevation, evapotranspiration, temperature, NDVI, slope); indicates the effect of amplification in the interaction of factors. However, the degree of contribution of individual groups of factors is described generically, without explicitly comparing their relative importance, which somewhat reduces the analytical depth.
3. Quantify the direct and indirect effects of factors on spatial NPP patterns using PLS-SEM. Partially implemented. The conclusion indicates the direction of direct and indirect effects; emphasizes the leading role of topography; describes the indirect effects of climate and anthropogenic factors. However: There are no quantitative estimates of the effects; the significance of the model is not discussed. This makes the contribution of PLS-SEM less pronounced compared to the stated goal.

In addition, I advise the authors to synthesize the conclusions by adding a final paragraph that will combine all three goals and clearly identify the scientific and practical contribution of the research.

The list of references must be designed according to the MDPI requirements.

Author Response

Comment1: Despite the large number of methods used, the methodology is not completely clear: how exactly are the results of trend analysis, Geodetector, and PLS-SEM logically related; are the results of some methods used as inputs or hypotheses for others? This gives the impression of parallel rather than hierarchically linked analysis.

Response1: Thank you for your insightful comment regarding the logical flow and interconnection of the methodologies employed in our study. We agree that clarifying the hierarchical relationship between the analytical methods is crucial for understanding the research design. The methods were not applied in parallel but were structured in a sequential, logical cascade where the outputs and insights from one stage informed the focus and setup of the next. The logical flow is: Descriptive Stats (What & Where are the changes?) → Trend Forecast (Will it continue?) → Driver Screening (What factors are correlated with it?) → Pathway Modeling (How do these factors work together causally?). This structured approach ensures that each methodological phase addresses a specific research question and provides the necessary foundation for the next, leading to a comprehensive understanding of NPP dynamics.We have revised the manuscript to explicitly state this logical sequence at the beginning of Section 2.3 and refined the introductory sentences for each subsection to reflect their role within this cascade. Thank you for prompting this clarification, which we believe significantly strengthens the methodological presentation.

Comment2: Insufficient elaboration of the PLS-SEM methodology.Although the conceptual model has been described, there is no information on checking the reliability and validity of measurement models; model quality criteria. This reduces the strictness of the interpretation of cause-and-effect relationships.

Response2: We have added a dedicated subsection (2.2.2.4. ) detailing the assessment of the measurement model's reliability/validity and the structural model's quality criteria (including R², Q², and GOF with explicit thresholds). Please see the updated manuscript and new Table 3 for the complete evaluation results.

 

Comment3: Insufficient description of data sources and quality. The methodology lacks a detailed discussion of input data uncertainties, possible remote sensing errors, and differences between climate and socio-economic data sources.

Response3: We have added a dedicated subsection "Data Uncertainties and Limitations"​ to explicitly address scale mismatches, algorithm biases in remote sensing products, limitations of socioeconomic proxies, and data interpolation uncertainties. This addition clarifies how these factors propagate through the analysis and provides context for interpreting the results.

Comment4: Nevertheless, despite its methodological completeness, the conclusion does not fully reflect the applied purpose of the study related to the assessment of the effectiveness of environmental remediation measures and their contribution to carbon neutrality strategies, which were clearly outlined in the formulation of the goal.

Response4: We have incorporated this suggestion and revised the conclusion section accordingly.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study analyzes the spatiotemporal evolution and driving mechanisms of Net Primary Productivity (NPP) in Hainan Island from 2001 to 2022 using MODIS data. Methods such as Trend Analysis and the Geographical Detector are employed. The long time span and comprehensive analytical framework provide reference value for ecological monitoring of tropical islands.

1. (Page 2, Lines 65-79): Hainan's tropical monsoon climate results in frequent cloud cover, which can limit the quality of optical remote sensing products. The authors directly use the MOD17A3HGF product without mentioning local accuracy validation or citing relevant regional validation literature. This may introduce bias into the results. It is recommended to supplement this with relevant validation studies or a local accuracy assessment.

2. (Page 20, Lines 563-587): Typhoons frequently make landfall in Hainan and significantly impact forest ecosystems and NPP. The analysis relies solely on annual average temperature and precipitation, overlooking the driving role of extreme weather events. It is suggested to add an analysis of typhoon disturbance in the discussion, selecting classic typhoon years to examine NPP changes before and after their passage.

3. (Page 20, Lines 566-567): Changes in NPP often exhibit a lag in response to climatic factors, which is not addressed. It is recommended to discuss this lag effect in the vegetation-climate relationship. For greater rigor, consider performing a correlation analysis between NPP and climate factors from the preceding period.

4. The authors use the Hurst exponent to predict future trends, noting that areas with a degradation trend and Hurst > 0.5 may continue to deteriorate. It is suggested to create a future zoning map based on the Hurst exponent and change slope, identifying areas of: ① sustained improvement, ② sustained degradation, and ③ improvement followed by degradation, accompanied by corresponding management recommendations.

5. While the abstract and introduction highlight human activity as a key factor, the study lacks the residual analysis commonly used to disentangle climatic and anthropogenic drivers of NPP. It is recommended to supplement this with a residual analysis or, in the discussion, clearly acknowledge the limitations of the employed method (Geographical Detector) in quantifying the contribution rate of human activities.

Author Response

Comment 1: (Page 2, Lines 65-79): Hainan's tropical monsoon climate results in frequent cloud cover, which can limit the quality of optical remote sensing products. The authors directly use the MOD17A3HGF product without mentioning local accuracy validation or citing relevant regional validation literature. This may introduce bias into the results. It is recommended to supplement this with relevant validation studies or a local accuracy assessment.

Response 1: We thank the reviewer for this important methodological point. We agree that cloud cover is a significant source of uncertainty for optical remote sensing products in tropical regions. In response, we have added a dedicated paragraph in Section 4.4 "Data Uncertainties and Limitations" to explicitly acknowledge this limitation and cite relevant regional validation studies. The added text clarifies that while cloud contamination may introduce biases in absolute NPP values, the MOD17A3HGF annual product has been validated in similar humid subtropical regions and is considered robust for analyzing relativespatial patterns and interannual trends—which is the primary focus of our study. We also suggest future validation work using local flux tower data (e.g., from the Jianfengling station). Please refer to the updated Section 4.4, Paragraph 1 for details.

 

Comment 2: (Page 20, Lines 563-587): Typhoons frequently make landfall in Hainan and significantly impact forest ecosystems and NPP. The analysis relies solely on annual average temperature and precipitation, overlooking the driving role of extreme weather events. It is suggested to add an analysis of typhoon disturbance in the discussion, selecting classic typhoon years to examine NPP changes before and after their passage.

Response 2: We appreciate this valuable suggestion. In the revised manuscript, we have expanded the discussion on the role of extreme weather events, particularly typhoons. Specifically, in Section 4.1 "Analysis of the spatiotemporal variation trends of NPP", we now explicitly discuss the potential link between notable low-NPP years (e.g., 2005) and major typhoon events. We acknowledge that our use of annual data integrates such transient impacts but does not isolate them. Furthermore, in Section 4.3.1 "Spatial heterogeneity", we have added a paragraph stating that our driver analysis, based on annual climate averages, may smooth out the intense but short-term impacts of typhoons. We suggest that future research incorporating extreme climate indices or event-based analysis could better elucidate these effects. This addition highlights the importance of extreme events for a comprehensive understanding of NPP dynamics in Hainan.

 

Comment 3: (Page 20, Lines 566-567): Changes in NPP often exhibit a lag in response to climatic factors, which is not addressed. It is recommended to discuss this lag effect in the vegetation-climate relationship. For greater rigor, consider performing a correlation analysis between NPP and climate factors from the preceding period.

Response 3: We thank the reviewer for raising this valid point regarding potential lag effects. Our study was primarily designed to identify spatialdrivers of NPP heterogeneity using methods (Geographical Detector and PLS-SEM) focused on spatial attribution within a given temporal unit (annual aggregates). Introducing lagged temporal analysis would require a different methodological framework (e.g., cross-correlation with higher-resolution data) and shift the focus from our core spatial-explicit objective. We acknowledge this as a limitation of the current study. In the revised manuscript, we have added a statement in the Limitations section (Section 4.4) to explicitly discuss this point, noting that NPP may respond to climate conditions from preceding seasons and suggesting that future work with higher temporal resolution data could investigate these temporal dynamics.

 

Comment 4: The authors use the Hurst exponent to predict future trends, noting that areas with a degradation trend and Hurst > 0.5 may continue to deteriorate. It is suggested to create a future zoning map based on the Hurst exponent and change slope, identifying areas of: ① sustained improvement, ② sustained degradation, and ③ improvement followed by degradation, accompanied by corresponding management recommendations.

Response 4: We thank the reviewer for this constructive suggestion. Indeed, our analysis using the Hurst exponent and Sen's slope allows for such a classification. In the revised manuscript, we have clarified this outcome. Figure A2 (Appendix) now explicitly presents a future trend zoning map based on the combination of the historical trend (K) and the Hurst exponent (H), categorizing areas into: (1) Persistent Increase, (2) Persistent Decrease, (3) Increase to Decrease (Anti-persistent), and (4) Decrease to Increase (Anti-persistent). Furthermore, in Section 4.2 "Analysis of the sustainability characteristics of NPP", we have structured the discussion and management recommendations specifically around these four identified future trend categories, providing targeted suggestions for each zone.

 

Comment 5: While the abstract and introduction highlight human activity as a key factor, the study lacks the residual analysis commonly used to disentangle climatic and anthropogenic drivers of NPP. It is recommended to supplement this with a residual analysis or, in the discussion, clearly acknowledge the limitations of the employed method (Geographical Detector) in quantifying the contribution rate of human activities.

Response 5: We agree with the reviewer that clearly stating the methodological boundaries is important. While residual analysis is a valuable method, our study aimed to employ a spatially explicit, multi-factor interaction framework (Geographical Detector) and a causal pathway model (PLS-SEM) to assess drivers. To address the concern about quantifying human activity contributions, we have added a specific paragraph in Section 4.3.1 "Spatial heterogeneity" to discuss the limitations of the Geographical Detector method in this context. We explicitly state that the method's reliance on proxy indicators (like nighttime lights) and its basis on spatial association mean that the calculated q-values for human activities should be interpreted as relative measures within our specific data framework, not as precise quantitative contribution rates. This addition enhances the transparency regarding the interpretation of our results concerning anthropogenic drivers.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The authors of the manuscript address a problem of great relevance in the field of ecology and sustainable management of natural resources, namely the temporal and spatial evolution of Net Primary Vegetation Productivity (NPP) and the factors influencing it on Hainan Island between 2001 and 2022. The study is important given the role of NPP in the carbon cycle and its contribution to the sustainability of tropical ecosystems, and it fills a clear gap in the quantitative and comprehensive analysis of NPP in tropical regions. However, we suggest that the authors update the data to include the years 2024 and 2025. Furthermore, we suggest reducing the similarity index, respectively from 38%, to below 10%.

The manuscript stands out for its interdisciplinary approach, combining statistical, geostatistical, and structural modeling methods to explore both historical trends and the determinants of NPP. The use of methods such as the Coefficient of Variation, Theil-Median trend, Mann-Kendall test, future trends analysis, Geographic Detectors, and PLS-SEM allow for a robust and integrated examination of the phenomenon. Nevertheless, to enhance clarity and reproducibility, we recommend reducing the similarity index from 38% to below 10% by rephrasing sections with close wording to cited sources and by synthesizing methodological and conceptual explanations.

The methodology is appropriate for the objectives, and the integration of various determinants provides a systemic perspective on the mechanisms influencing NPP. However, clarification of the criteria for data selection and processing is recommended, as well as highlighting limitations related to modeling and the generalizability of results to other tropical regions.

The results are presented both descriptively and graphically and are clear and relevant. For example, NPP showed an average annual increase of 3.6 g C·m²·yr⁻¹, with a decreasing distribution from the central south toward the coastal areas; future trends indicate a predominance of “continuous decrease” and “increase followed by decrease” scenarios. Topographic, climatic, anthropogenic, and vegetation-related factors influence NPP, with human activities having indirect effects by interacting with latent variables that facilitate vegetation growth.

In conclusion, the manuscript makes a valuable theoretical and practical contribution, providing relevant data for land use planning, afforestation programs, and ecological protection and restoration measures. We recommend that the authors more explicitly highlight the study’s original contribution, clarify methodological limitations, and reduce the similarity index to ensure the manuscript’s uniqueness and scientific quality.

We congratulate the authors for their innovative approach, practical applicability, and ecological relevance of the results, and we recommend revising the manuscript to strengthen the emphasis on its scientific and applied impact, particularly in the abstract, results, and discussion sections.

Author Response

Comment 1: We suggest that the authors update the data to include the years 2024 and 2025. Furthermore, we suggest reducing the similarity index, respectively from 38%, to below 10%.

Response 1: We thank the reviewer for the suggestions regarding data currency and manuscript originality.

Data Update to 2024/2025: The MOD17A3HGF NPP product serves as the core dataset for our analysis. However, the data for the driving factors include climatic factors (evapotranspiration, rainfall, and temperature), topographical factors (aspect, elevation, and slope), vegetation factors (Normalized Difference Vegetation Index, NDVI), and anthropogenic activities (Gross Domestic Product (GDP), population, nightlights, and land use (LUCC)). To ensure the consistency and reliability of the entire time-series analysis across all datasets, we maintained the 2001-2022 study period. We acknowledge the value of incorporating the most recent data and have added a note in the "Data Uncertainties and Limitations"​ section (Section 4.4) stating that future work will benefit from including the latest available data to update the trend analysis as newer validated product versions are released.

Reducing Similarity Index:​ We have taken comprehensive steps to reduce textual similarity. The original percentage was primarily elevated due to the standard technical descriptions of well-established methodologies (e.g., Theil-Sen, Mann-Kendall, coefficient of variation). We have:Thoroughly paraphrased the methodological descriptions in Sections 2.3.1 to 2.3.6, focusing on explaining the application and rationale in the context of our specific study rather than providing generic definitions.Significantly rewritten the Introduction​ and Discussion​ sections to present arguments and interpretations in our own words, supported by recent and relevant citations.

 

Comment 2: However, clarification of the criteria for data selection and processing is recommended, as well as highlighting limitations related to modeling and the generalizability of results to other tropical regions.

Response 2: We thank the reviewer for this suggestion to enhance methodological transparency and contextualize our findings.

Clarification of Data Selection and Processing:​We have expanded Section 2.2 "Data Sources and Methods"​ to provide a clearer rationale for data source selection. Furthermore, we have added a more detailed subsection, and Table 1​ specifically describes each dataset to ensure reproducibility.

Highlighting Modeling Limitations and Generalizability:​We have strengthened the "Data Uncertainties and Limitations"​ section.

 

Comment 3: We congratulate the authors for their innovative approach, practical applicability, and ecological relevance of the results, and we recommend revising the manuscript to strengthen the emphasis on its scientific and applied impact, particularly in the abstract, results, and discussion sections.

Response 3: We are grateful for the positive assessment and have revised the manuscript throughout to better highlight its contributions.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The article is titled ‘Temporal and spatial evolution of vegetation net primary productivity and key contributing factors on Hainan Island from 2001 to 2022’. The aim of this study is to analyze the temporal and spatial evolution of vegetation net primary productivity. The study was conducted for Hainan Island from 2001 to 2022.

My comments are as follows:

- the introduction needs to be expanded – what research gap does this study fill?

- the study area should include a figure with the study area location.

- the methodology section requires substantial improvement. Sections 2.2 and 2.3 should be combined as data and methods. The section on methods should be, for example, 2.2.1. Tables with data sources should be included in this section. Subsection 2.3, methodology, should include a research flowchart. It should be clear to the reader what the order of the studies was, and then formulas and indicators should be provided. The layout of the paper in its current form is unacceptable.

- The discussion requires improvement; the authors should refer to the literature.

- The conclusions require improvement; these should be the main findings of the study. Technical Notes :

- Abbreviations in keywords should be avoided.

- The list of references should be adapted to the journal's requirements.

- The layout of the paper requires improvement. This applies to the naming and division of content for sections and subsections, e.g., '3.1. Subsection' – a specific title is missing; a separate section can refer to lines 277, 290, or 339. The subsection cannot be titled '3.2. Figures, Tables, and Schemes'.

- Literature citation needs improvement – ​​see lines 565 and 569.

 

Author Response

Comment 1: the introduction needs to be expanded – what research gap does this study fill?

Response 1: The research gap has been addressed in the Introduction section.

 

Comment 2:the study area should include a figure with the study area location.

Response 2: We agree that a study area map is essential for providing clear geographical context. In response, we have added a new figure illustrating the location of Hainan Island. This map, now included in the revised manuscript, highlights the island's position within China and provides key geographical references.Please see the new Figure 1: Overview of Hainan Island.

 

Comment3: the methodology section requires substantial improvement. Sections 2.2 and 2.3 should be combined as data and methods. The section on methods should be, for example, 2.2.1. Tables with data sources should be included in this section. Subsection 2.3, methodology, should include a research flowchart. It should be clear to the reader what the order of the studies was, and then formulas and indicators should be provided. The layout of the paper in its current form is unacceptable.

Response3: Agreed to the modifications. Sections 2.2 and 2.3 have been merged into the "Data and Methods" section, and a research flowchart has been added to the methodology subsection (2.2.2).Please see the new Figure 2:Study’s framework.

 

Comment4:The discussion requires improvement; the authors should refer to the literature.

Response4: We sincerely thank the reviewer for this valuable suggestion. In response, we have thoroughly revised the Discussion section to strengthen its foundation in the existing literature.Please refer to the revised Sections 4.1, 4.2, 4.3, and 4.4 for detailed changes. We believe these revisions have significantly enhanced the scholarly depth and credibility of the discussion.

 

Comment5:The conclusions require improvement; these should be the main findings of the study. Technical Notes :

Response5:We thank the reviewer for this constructive suggestion. In response, we have revised the Conclusion​ section to ensure it concisely and clearly highlights the main findings of the study. The revised conclusion now focuses on summarizing the key results regarding the spatiotemporal trends, future projections, dominant drivers, and their interrelationships, as derived from our integrated analytical framework. This modification strengthens the link between the core findings and their scientific implications.

 

Comment6:Abbreviations in keywords should be avoided.

Response6:We thank the reviewer for this correction. All abbreviations in the Keywords section have been replaced with their full terms in the revised manuscript.

 

Comment7:The list of references should be adapted to the journal's requirements.

Response7:We confirm that the reference list has been thoroughly revised and formatted according to the journal’s author guidelines in the updated manuscript.

 

Comment8:The layout of the paper requires improvement. This applies to the naming and division of content for sections and subsections, e.g., '3.1. Subsection' – a specific title is missing; a separate section can refer to lines 277, 290, or 339. The subsection cannot be titled '3.2. Figures, Tables, and Schemes'.

Response8:We agree with this comment and have revised the manuscript structure accordingly. All section and subsection headings have been updated with specific, descriptive titles. Content previously listed under generic headings (e.g., "3.2. Figures, Tables, and Schemes") has been integrated into its respective logical sections within the main text. Please refer to the revised manuscript for the updated layout.

 

Comment9:Literature citation needs improvement – ​see lines 565 and 569.

Response9:We sincerely thank the reviewer for pointing out the need for improvement in literature citation format. In response, we have carefully reviewed and revised all in-text citations throughout the manuscript to ensure they strictly adhere to the journal's formatting guidelines. Specifically, we have corrected the citation style in lines 565 and 569, as well as in other relevant sections, by properly formatting them as superscripts. All references have been verified and consistently formatted in the reference list accordingly.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have significantly improved the paper. I have no further comments.

Author Response

Comprehensive revisions have been made to the entire text, including adding a roadmap and improving the method introduction in the methods section, as well as adding the existing limitations and the next research direction in the results and discussion section. For details, please refer to the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

After revision, the paper has been significantly improved, but a similarity rate of 34% has been detected, suggesting the need for plagiarism reduction. Additionally, it is advisable to self-check whether the paper contains an excessively high proportion of AI-generated content.

Author Response

The plagiarism rate has dropped to 8%.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Editor,

Many thanks for your cooperation!

We also appreciate the revision carried out in accordance with the recommendations provided. However, in the Conclusions section, we suggest highlighting directions for future research, especially in the context of the limitations outlined in Section 4.4. Thank you!

I wish you and your team all the best!

Kind regards,

Otilia MANTA

Author Response

The 4.4 section has been modified as required. The modifications are as follows:

4.4 Limitations and Future Work​

The data collected in this study were primarily sourced from the GEE platform. Using these datasets and relevant models, we analyzed the key factors influencing vegetation NPP variations in Hainan Province and their underlying mechanisms. We further examined the spatiotemporal patterns of these changes to identify correlations between NPP and major influencing factors. The PLS-SEM model was employed to quantify the impact levels of different factors. The findings provide valuable insights for ecological management and land use policy formulation in the study area, while also offering references for similar research. However, several limitations remain. Firstly, the applicability of the MOD17A3HGF NPP product to humid subtropical regions remains unclear, and the interannual variation results of calculated NPP exhibit relatively low accuracy.^{[64, 65]}. When applied to calculate net primary production (NPP) under tropical climate conditions in Hainan, the model's results may deviate from actual conditions, compromising the reliability of conclusions. To effectively address this issue, it is essential to collect more field measurements and regional calibration data across different areas, thereby establishing a tailored CASA model that better meets local NPP analysis requirements. Secondly, while this study employs high-temporal-resolution remote sensing data for model analysis, other datasets still lack sufficient resolution, limiting the model's ability to analyze small-scale NPP variations. Furthermore, the results fail to account for factors such as small-scale land use changes and short-term extreme weather events. Therefore, comprehensive analysis incorporating additional resource survey data is crucial to obtain more reliable outcomes. The model's interactions among multiple variables remain underexplored. For complex tropical forest ecosystems, these factors exhibit intricate synergistic effects and sometimes undergo dramatic nonlinear changes, necessitating future research to analyze these dynamics^[66]. In addition, the influence of socio-economic factors on NPP is also obvious. In the future, it is necessary to analyze the interaction between these factors and ecosystem productivity, and determine the influence of each major factor on NPP, so as to obtain more reliable results.

Reviewer 4 Report

Comments and Suggestions for Authors

The authors only partially revised the article. The article still requires improvement.

- Subsection 2.2.2. was described very generally. Each step of the procedure should be described here.

- The conclusions require improvement. The main findings resulting from the analysis should be indicated. The text should not be divided into two subsections.

Technically, the article is poorly prepared. The authors declared that they made the revisions very carefully, which is contradicted by this manuscript.

- The literature citation method is inconsistent with the journal's standards – I suggest you see other articles.

- The figure showing the study area is of little use; a map showing land use and the river network should be added.

- The list of references requires improvement. The list should include all authors of the article, not just the first author.

- Tables should be adapted to the journal's requirements.

Author Response

Please refer to the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 4 Report

Comments and Suggestions for Authors The authors have partially revised the article. The article still requires extensive editorial improvement in accordance with the journal's requirements, such as the manner of citing literature in the text. I leave this matter to the journal's editor.      

Author Response

Thank you for your guidance. We have carefully revised the manuscript in accordance with the editorial requirements, rephrasing the duplicate text (highlighted in red in the manuscript), with particular attention to the sections marked as primary sources in the similarity report. We will further improve the text formatting as requested in the subsequent steps.