Geospatial Explainable AI Uncovers Eco-Environmental Effects and Its Driving Mechanisms—Evidence from the Poyang Lake Region, China

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article utilizes the Geospatial Explainable Artificial Intelligence (GeoXAI) framework to study the ecological and environmental effects of land-use transitions, which holds significant value. The research topic is highly relevant to current needs, with rigorous methods and a clear structure. The following are the revision suggestions:

1. In the abstract, specific innovative aspects of the study could be appropriately added. For example, explain how GeoXAI addresses the deficiencies of traditional methods and the potential application value of the research results in other regions.
2. In the introduction, when elaborating on the research questions, more specific research hypotheses should be clearly stated to enable readers to better understand the direction of the study.
3. Regarding the GeoXAI framework, in addition to comparing its performance with traditional methods, its theoretical basis and innovative principles should be further elaborated to help readers better understand the scientific nature of this method.
4. In the result analysis, the interpretation of the results should be strengthened. Explain the ecological significance behind each result. For example, when describing the impact of different factors on the EQI, relevant ecological processes can be incorporated for illustration.
5. In the discussion section, the author could engage more deeply with previous studies. Instead of merely summarizing existing research, a more critical analysis of the similarities and differences between the research results and prior work should be conducted.
6. Currently, the conclusions mainly focus on the Poyang Lake region. To enhance the generalizability of the research results, the author could discuss how these results can be applied to other regions with similar ecological and socio-economic characteristics.
7. A prospect of future research could be added. For example, how to further improve the GeoXAI framework and how to apply this method to the study of other ecosystems.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study proposes a GeoXAI framework integrating geospatial machine learning and Shapley values to uncover the spatial heterogeneity of driving mechanisms of eco-environmental quality. The paper is well-organized and presents several interesting findings. The figures are clear and easy to interpret. However, the authors should refine ambiguous expressions and further clarify the methods to avoid potential misunderstandings.

Detailed comments:

Title

The current title may lead to misunderstandings, as it does not clearly convey the specific relationship the authors aim to analyze. Is it the relationship between land use transitions and eco-environmental quality? The relationship between driving mechanisms and land use transitions? Or the relationship between driving mechanisms and eco-environmental quality? If this study focused on the third relationship, how did the authors distinguish the effects of land use transitions from other driving factors on eco-environmental quality? Besides, the abstract should be reorganized to clarify these issues clearly.

Abstract

1. Although the keywords included “Production-Living-Ecological space”, the abstract does not appear to address this concept. The abstract should clearly reflect and elaborate on the listed keywords.
2. Add the key findings on the eco-environmental effects of land use transitions.

Introduction

1. Lines 50-51: “The World Bank estimates that environmental issues in China result in annual economic losses amounting to 10% of the country’s GDP”, which environmental issues? And how exactly do these issues impact China’s economy? Please clarify.
2. Lines 57-58: “Accordingly, optimizing and refining the spatial development pattern has become a central strategy in advancing ecological civilization.” Is this study focused on optimizing spatial development patterns? If not, consider revising the sentence to better align with the paper’s actual objectives.
3. Lines 74-76: “Further investigations into the impacts and underlying mechanisms 74 of land-use transformation across various regional scales are essential for informing land-75 use management.” Please further explain how this point connects to this study.
4. Lines 78-99: The authors cite research that used GeoDetector to highlight limitations of spatial statistical methods in capturing nonlinear relationships. However, GeoDetector does not rely on a linearity assumption, and it can identify interactions of different influencing factors.
5. Lines 96-111: The distinction between GeoXAI and the Random Forest model should be clarified, especially in terms of their “ability to capture the complex nonlinear relationships”.
6. To the best of my knowledge, GeoXAI has already been applied in several geographical studies. The authors are suggested to investigate whether existing literature has used GeoXAI to analyze eco-environmental quality.

Materials and Methods

1. Line 155: “the Eco-environmental quality of the study area was divided into 7243 hexagonal evaluation units, each measuring 1 km × 1 km.” Please clearly clarify this statement. Besides, at what spatial scale was the EQI calculated and interpolated?
2. Line 156: “This division was achieved through iterative comparisons across different spatial scales.” By what criteria?
3. Table 2: Add the spatial resolution and time periods of the data, either within the table or in the main text.
4. Formula (2): How was Ri determined? And how to decide the EQI value in Table 1?
5. Line 244: GE0 should be GEO.
6. Figure 2: The relationship between “Ecological Contribution Rate” and EQI is unclear., The calculation of EQI in section 2.3.2 does not seem to be directly related to Ecological Contribution Rate. Furthermore, what is the meaning of X1, X2, X3?

Results, Discussion, and Conclusions

1. Lines 350-351: Please clarify why avoiding multicollinearity is necessary when using a MECHAINE LEARNING approach?
2. Figure 3: The font size in the figure (especially the chord diagram) should be enlarged.
3. Line 365: Clarify the meaning of GEO.
4. Line 565: In the CONCLUSIONS section, replace “PopDens” with its full descriptive term for clarity.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I have attached my review document to this section.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed my concerns well, and I recommend accepting the paper.

Reviewer 3 Report

Comments and Suggestions for Authors

The revised draft has exceeded my expectations. I appreciate the authors' dedication. I have no further comments and hope many readers will engage with your paper.