Spatial Correlation Network Characteristics of Comprehensive Transportation Green Efficiency in China

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper evaluates China's CTGE using the SBM-DEA, SDE, SNA models. This is a crucial issue; however, the paper requires improvements in its academic contribution.

1. The abstract lacks an explanation of the DEA model, and its numerical results. Additionally, the study’s academic and practical contributions should be emphasized.
2. The introduction and literature review sections should be clearly distinguished. The introduction should explain the background and significance of the research problem, while the literature review should systematically summarize related academic research trends.
3. Previous studies have used the DEA model to evaluate sustainable transportation service quality. It is necessary to include more relevant and recent research to ensure a state-of-the-art review: https://doi.org/10.1016/j.retrec.2024.101491; https://doi.org/10.1155/2018/6701484
4. The methodology section explains the SBM-DEA, SDE, and SNA models individually; however, the overall research framework remains unclear. A new subheading should be added at the beginning of the methodology section to define the research problem and provide a comprehensive summary of how each methodology is applied.
5. It is necessary to justify the use of the three methodologies. In fact, the entire process could potentially be conducted using only the DEA model.
6. The data description is missing. A section should be added at the beginning of the Results and Discussion section to describe the data used in this study. Additionally, presenting the basic statistical characteristics of the dataset in a table could improve readability.
7. The efficiency scores derived from the DEA model are a key result of the study and require a more detailed explanation. Currently, efficiency scores are only briefly mentioned, and specific numerical findings are insufficient. Therefore, a table summarizing the efficiency scores should be included, along with a more in-depth discussion of the results.
8. The structure of the results section needs to be reorganized. Subheadings should be structured according to the methodological flow (SBM-DEA → SDE → SNA), and each section should be logically structured.
9. Since the number of DMUs analyzed is only 30, using the SBM model may not be necessary. If the results do not differ significantly when compared with the VRS (BCC) or CRS (CCR) models, a simpler model should be considered.
10. One of the major advantages of the DEA model is that it provides benchmarking information for inefficient DMUs. Typically, DEA studies derive improvement strategies for inefficient regions using benchmarking DMUs. It is necessary to compare benchmarking-based strategies with SNA-based strategies to assess their effectiveness.
11. There are numerous typographical errors in the manuscript. A thorough review should be conducted to correct them, and the consistency and logical flow of the sentences should also be checked.

Author Response

This paper evaluates China's CTGE using the SBM-DEA, SDE, SNA models. This is a crucial issue; however, the paper requires improvements in its academic contribution.

Response: Sincere thanks for your insightful and accurate comments on our manuscript. We have benefited a lot from your valuable suggestions and comments, on the basis of which the manuscript has been thoroughly revised.

Comments:

1. The abstract lacks an explanation of the DEA model, and its numerical results. Additionally, the study’s academic and practical contributions should be emphasized.

Response: We thank the reviewer for this insightful suggestion. We agree with this comment. Based on your recommendations, we have rewritten the abstract with enhanced descriptions of the DEA methodology and relevant datasets, while further clarifying the academic and practical contributions of this research. For details, please refer to lines 11-34 on page one of the manuscript.

2. The introduction and literature review sections should be clearly distinguished. The introduction should explain the background and significance of the research problem, while the literature review should systematically summarize related academic research trends.

Response: We are thankful to your valuable suggestion. We restructured Section 1:Introduction (Section 1) now focuses on the research background, problem significance, and policy context (Lines 38–112). A new Literature Review subsection (Section 2) systematically summarizes prior studies on transportation efficiency, methodological trends, and research gaps (Lines 113–196).

3.Previous studies have used the DEA model to evaluate sustainable transportation service quality. It is necessary to include more relevant and recent research to ensure a state-of-the-art review: https://doi.org/10.1016/j.retrec.2024.101491; https://doi.org/10.1155/2018/6701484

Response: We are appreciated for your professional comments. After reviewing the literature, we cited more recent and pertinent research. For example:

Eun Hak Lee. eXplainable DEA approach for evaluating performance of public transport origin-destination pairs[J]. Research in Transportation Economics, 2024, 108: 101491. https://doi.org/10.1016/j.retrec.2024.101491

Guo J, Nakamura F, Li Q, et al. Efficiency Assessment of Transit-Oriented Development by Data Envelopment Analysis: Case Study on the Den-en Toshi Line in Japan [J]. Journal of Advanced Transportation, 2018, 6701484. https://doi.org/10.1155/2018/6701484.

Guo, J., Bai, S., Li, X., Xian, K., Liu, E., Ding, W., & Ma, X. A universal geography neural network for mobility flow prediction in planning scenarios. Computer‐Aided Civil and Infrastructure Engineering. 2025, 13398. https://doi.org/10.1111/mice.13398.

Li, A., Xu, Z., Li, W., Chen, Y., & Pan, Y. . Urban Signalized Intersection Traffic State Prediction: A Spatial–Temporal Graph Model Integrating the Cell Transmission Model and Transformer. Applied Sciences, 2025, 15(5), 2377. https://doi.org/10.3390/app15052377.

Li, S J., Ma, Q F. Carbon quota allocation and emission reduction responsibility sharing at provincial level in China from transport industry [J]. Sustainable Futures, 2025, 9: 100535. https://doi.org/10.1016/j.sftr.2025.100535.

Ma, H. T. Urbanization under globalization: How does the Belt and Road Initiative affect urbanization levels in participating countries [J]. Journal of Geographical Sciences, 2022, 32 (11): 2170-2188. https://doi.org/10.1007/s11442-022-2042-1.

Cao, Y. W., Zhang, R. R., Zhang, D. H., et al. Urban Agglomerations in China: Characteristics and Influencing Factors of Population Agglomeration[J]. Chinese Geographical Science, 2023, 33(04): 719-735. https://doi.org/10.1007/s11769-023-1368-7.

Upadhyay, R. K., Sharma, S. K., Kumar, V. Introduction to Intelligent Transportation System and Advanced Technology[J]. Energy, Environment, and Sustainability, Springer, Singapore, 2024. https://doi.org/10.1007/978-981-97-0515-3_1.

4. The methodology section explains the SBM-DEA, SDE, and SNA models individually; however, the overall research framework remains unclear. A new subheading should be added at the beginning of the methodology section to define the research problem and provide a comprehensive summary of how each methodology is applied.

Response: Thank you for your advice. According to your advice, we created a research framework at the beginning of section 3, expanded the specific problems to be handled by each approach in detail, and displayed them in a graphic fashion to help readers' understanding. Please see page 5, line 198-210.

“3.1 Research framework

The research framework is illustrated in Figure 1. First, the SBM-DEA model is employed to compute the CTGE (Comprehensive Transportation Green Efficiency) for 30 Chinese regions from 2003 to 2022 by inputting relevant indicators, followed by an analysis of its temporal evolution trends and spatial distribution patterns. Second, the Standard Deviational Ellipse (SDE) model is applied to identify the dynamic evolution characteristics of CTGE’s spatial configuration, including the movement trajectory of the center of gravity and the parametric characteristics of the standard deviational ellipse. Finally, spatial linkages between regional CTGE values are identified using the Granger causality test, and a social network analysis (SNA) model is constructed to examine the global structural features, individual nodal characteristics, and inter-block interaction mechanisms within the CTGE spatial correlation network.”

5.It is necessary to justify the use of the three methodologies. In fact, the entire process could potentially be conducted using only the DEA model.

Response: Thank you for your suggestion. Based on your recommendations, we have elaborated in detail on the roles of relevant research methods and the rationale for selecting them in the research framework and methodology section. This study primarily encompasses three research components:

First, the SBM-DEA model is employed to scientifically measure the CTGE (Comprehensive Transportation Green Efficiency) of 30 regions in China. According to previous research, the DEA model is one of the most widely used and effective methods for efficiency measurement. The SBM model further incorporates undesirable outputs into the model without requiring consideration of indicator dimensionality, thereby enabling more accurate efficiency assessments. The measurement of CTGE in this study includes desirable outputs (economic growth and social progress) and undesirable outputs (carbon emissions from the transportation sector), which align with the data requirements of the model. Consequently, the SBM-DEA model is adopted to measure China's CTGE.

Second, the Gravity Center-Standard Deviation Ellipse model is utilized to evaluate the dynamic evolutionary trends and spatial distribution characteristics of China's CTGE. This method is commonly applied to analyze the spatial distribution patterns of research subjects, revealing the concentration and dispersion of data. For instance, the dynamic shifts of the gravity center reflect the temporal migration of the distribution center for geographical elements, the rotation angle of the standard deviation ellipse indicates directional changes in data distribution, and the ellipse area reflects the overall dispersion range of data. As transportation serves as an essential geographical element in production and daily life, accurately measuring the spatial pattern changes of CTGE can holistically reflect regional disparities and agglomeration effects across different areas, laying the foundation for identifying the network structural characteristics of CTGE in subsequent analyses.

Third, the Social Network Analysis (SNA) model is applied to investigate the spatial correlation network structure of CTGE among China's 30 regions. Social network theory is the most commonly used framework for constructing spatial correlation networks of elements, offering multiple quantitative metrics such as node degree, clustering coefficient, and centrality, which enable researchers to objectively evaluate nodes and relationships within the network. Simultaneously, the graphical representation and analysis of social networks allow researchers to intuitively understand network structures and features. As transportation serves as a carrier for the cross-regional flow of resource elements, the spatial correlations of CTGE across regions exhibit multi-threaded and complex network characteristics. Therefore, accurately identifying the spatial correlation network features of China's CTGE and uncovering the roles and positions of different regions within the network hold significant guidance for formulating targeted regional transportation policies and addressing regional disparities.

6. The data description is missing. A section should be added at the beginning of the Results and Discussion section to describe the data used in this study. Additionally, presenting the basic statistical characteristics of the dataset in a table could improve readability.

Response: Thank you for your valuable suggestions. Based on your recommendations, we have supplemented the data sources used in this study and included a table presenting the descriptive statistics of the relevant data in the Indicator selection and Data Sources to enhance the readability of the article. We sincerely appreciate your guidance once again. Please see page 10, line 411-435.

7. The efficiency scores derived from the DEA model are a key result of the study and require a more detailed explanation. Currently, efficiency scores are only briefly mentioned, and specific numerical findings are insufficient. Therefore, a table summarizing the efficiency scores should be included, along with a more in-depth discussion of the results.

Response: We are grateful for your constructive comments. Based on your recommendations, we have supplemented the CTGE calculation results (see Table 2) in Section 3.1 and expanded the discussion and analysis of these findings, providing a detailed explanation of the underlying and contributing factors behind the results. Please see page 11-14, line 447-504.

8. The structure of the results section needs to be reorganized. Subheadings should be structured according to the methodological flow (SBM-DEA → SDE → SNA), and each section should be logically structured.

Response: Thank you very much for your constructive suggestions. Based on your recommendations, we have revised the three subtitles in the Results Analysis section, optimizing their structure to follow the logical sequence of SBM-DEA → SDE → SNA. The updated subtitles are as follows:

4.1 Analysis of China's CTGE Measurement Results Based on the SBM-DEA Model

4.2. Spatiotemporal Evolution of China's CTGE Using the SDE Model

4.3. Network Correlation Characteristics of China's CTGE via the SNA Model

9.Since the number of DMUs analyzed is only 30, using the SBM model may not be necessary. If the results do not differ significantly when compared with the VRS (BCC) or CRS (CCR) models, a simpler model should be considered.

Response: Thank you for your feedback. First, the research subjects of this study are the CTGE (Comprehensive Transportation Green Efficiency) values of 30 regions in China from 2003 to 2020, resulting in 540 decision-making units (DMUs). This satisfies the DEA model’s requirement for the number of DMUs (i.e., the total DMUs should be at least three times the sum of input and output indicators). Second, the BCC model corresponds to the variable returns-to-scale (VRS) framework, while the CCR model aligns with constant returns-to-scale (CRS); both are radial models, whereas the SBM model is non-radial. The SBM model evaluates DMU efficiency by calculating slack variables in inputs and outputs. Compared to traditional CCR and BCC models, the SBM model more effectively addresses inefficiencies caused by input excesses (where actual inputs exceed theoretical minimums) or output shortfalls (where actual outputs fall below theoretical maximums). Finally, unlike conventional DEA models (CCR and BCC), the SBM model not only avoids biases introduced by radial or angular measurements but also incorporates undesirable outputs in the production process, thereby better reflecting the essence of efficiency evaluation. Based on these considerations, this study adopts the SBM-DEA model to measure China’s CTGE.

10.One of the major advantages of the DEA model is that it provides benchmarking information for inefficient DMUs. Typically, DEA studies derive improvement strategies for inefficient regions using benchmarking DMUs. It is necessary to compare benchmarking-based strategies with SNA-based strategies to assess their effectiveness.

Response: We sincerely appreciate the reviewer’s insightful suggestion. The comparison between DEA-based benchmarking strategies and SNA-driven strategies is indeed a valuable direction to enhance the practical implications of this study. In response, we have taken the following actions: Firstly, Methodological Clarification. Added a subsection in the Conclusions and policy implications (Section 5) to explicitly compare the two types of strategies: DEA Benchmarking: Focuses on identifying efficiency gaps for individual DMUs (e.g., reducing input redundancy or undesirable outputs) by referencing efficient frontiers. SNA-Driven Strategies: Prioritizes optimizing network roles (e.g., strengthening central nodes or bridging structural holes) to improve systemic coordination and resource flow. Secondly, Limitations and Future Work. Acknowledged in the Conclusion that the effectiveness of strategies may vary depending on regional resource endowments and network positions. Proposed future research to integrate DEA and SNA frameworks for hybrid policy design.

11.There are numerous typographical errors in the manuscript. A thorough review should be conducted to correct them, and the consistency and logical flow of the sentences should also be checked.

Response: Thank you very much for your valuable suggestions. In accordance with the journal’s guidelines, we have reformatted the manuscript and thoroughly reviewed all sentences to eliminate any logical inconsistencies. Additionally, the paper has undergone comprehensive proofreading and grammatical revisions by a native English speaker. We sincerely appreciate your constructive feedback once again.

Frankly, we appreciate this reviewer for insightful and accurate comments on our manuscript. Te be honest, we have spent a lot of effort and time improving this manuscript. We have carefully revised and responded to each comment raised by reviewer. Finally, we show great respect for reviewer’s professionalism. If we have any unclear explanations, please contact us immediately. Thank again for your positive and useful comments.

Yours Sincerely

The authors

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript explores the spatial correlation network characteristics of China's Comprehensive Transportation Green Efficiency (CTGE) using a combination of the Slack-Based Measure (SBM)-DEA, Social Network Analysis (SNA), and the Standard Deviation Ellipse (SDE) model. The study provides valuable insights into regional disparities in transportation efficiency, highlighting structural patterns and their implications for regional development policies.

The paper is well-structured and contributes to the existing literature on transportation efficiency by integrating network analysis techniques with spatial economic evaluation methods. However, several aspects require further elaboration and clarification:

1. Why was SBM-DEA chosen over other efficiency measurement techniques, such as the Directional Distance Function (DDF) or Stochastic Frontier Analysis (SFA)?
2. How robust are the results? Did the authors conduct sensitivity analyses to determine whether efficiency scores significantly change under different model specifications?
3. How were the input-output indicators selected? Is there empirical justification for including or excluding certain factors?
4. Granger causality primarily establishes temporal rather than spatial relationships. Have the authors considered spatial econometric techniques, such as Spatial Autoregressive (SAR) models or Geographically Weighted Regression (GWR), to confirm their findings?
5. The paper does not explicitly mention whether exogeneity concerns were addressed in the causality tests. Can the authors clarify how they ruled out omitted variable bias or reverse causality?
6. What are the economic and policy implications of being a "peripheral" province in the CTGE network? How can western and central regions improve their efficiency?
7. What role do regional policies, infrastructure investments, and technology diffusion play in shaping spillover effects? Can the authors provide empirical examples of successful interventions?
8. The authors identify four network blocks (main inflow, main outflow, bidirectional spillover, and agent block), but what specific policy recommendations apply to each block?
9. Why were these specific years chosen? Would including more recent years (e.g., post-pandemic data from 2021–2023) significantly alter the findings?
10. How reliable is the transportation industry data? Was any data imputation performed for missing values?
11. Are there sectoral variations in efficiency? Would a disaggregated analysis (e.g., road vs. rail vs. aviation) provide deeper insights?
12. Literature Review: The paper discusses existing research but would benefit from a more structured comparison of prior studies, particularly those that apply network analysis to transportation efficiency. It is recommended to include additional review papers:

Guo, J., Bai, S., Li, X., Xian, K., Liu, E., Ding, W., & Ma, X. (2025). A universal geography neural network for mobility flow prediction in planning scenarios. Computer‐Aided Civil and Infrastructure Engineering.

Li, A., Xu, Z., Li, W., Chen, Y., & Pan, Y. (2025). Urban Signalized Intersection Traffic State Prediction: A Spatial–Temporal Graph Model Integrating the Cell Transmission Model and Transformer. Applied Sciences, 15(5), 2377.

13. Please check the name of "Shaanxi" in Figure 4.

14. English expressions do not use full stops in certain contexts, so the symbol in Figure 6 should be a comma instead of a Chinese character.

15. Please check the entire manuscript for grammatical accuracy.

 

Comments on the Quality of English Language

Please check the entire manuscript for grammatical accuracy.

Author Response

The manuscript explores the spatial correlation network characteristics of China's Comprehensive Transportation Green Efficiency (CTGE) using a combination of the Slack-Based Measure (SBM)-DEA, Social Network Analysis (SNA), and the Standard Deviation Ellipse (SDE) model. The study provides valuable insights into regional disparities in transportation efficiency, highlighting structural patterns and their implications for regional development policies. The paper is well-structured and contributes to the existing literature on transportation efficiency by integrating network analysis techniques with spatial economic evaluation methods. However, several aspects require further elaboration and clarification.

Response: We sincerely thank the reviewers for their positive assessment of our manuscript and their insightful and constructive comments, which have significantly contributed to improving the quality of our work.

Comments:

1.Why was SBM-DEA chosen over other efficiency measurement techniques, such as the Directional Distance Function (DDF) or Stochastic Frontier Analysis (SFA)?

Response: We sincerely appreciate the reviewer’s insightful question regarding the selection of the SBM-DEA model. Below, we clarify the rationale for choosing SBM-DEA over alternative methods such as the Directional Distance Function (DDF) and Stochastic Frontier Analysis (SFA):

Non-Radial Efficiency Measurement. Unlike radial models (e.g., DDF), which assume proportional adjustments to inputs/outputs and may overestimate efficiency by ignoring slack variables, the SBM-DEA model explicitly incorporates input excesses and output shortfalls into efficiency calculations. This non-radial approach avoids the "proportionality bias" inherent in DDF, making it more suitable for evaluating transportation systems where resource allocation inefficiencies (e.g., redundant infrastructure or underutilized capacity) are critical.

Handling Undesirable Outputs. While both SBM-DEA and DDF can integrate undesirable outputs (e.g., carbon emissions), SBM-DEA directly penalizes such outputs through its objective function, ensuring that efficiency scores reflect environmental constraints without requiring pre-defined directional vectors. In contrast, DDF relies on subjective assumptions about adjustment directions (e.g., maximizing desirable outputs while minimizing undesirable ones), which may not align with the multi-objective nature of green transportation systems.

Non-Parametric Flexibility vs. Parametric Limitations. Compared to SFA, which requires specifying a production function and distributional assumptions for inefficiency terms, SBM-DEA is non-parametric and avoids functional form restrictions. This flexibility is particularly advantageous in our study, where the complex interdependencies among transportation modes (e.g., road-rail-waterway synergies) make it challenging to define a universally valid production function. Additionally, DEA’s deterministic nature aligns better with our policy-oriented goal of identifying specific improvement benchmarks, whereas SFA’s stochastic framework focuses more on statistical inference.

Empirical Consistency with Data Characteristics. Our dataset includes 540 decision-making units (DMUs) across 30 regions over 18 years, satisfying DEA’s requirement that DMUs exceed three times the sum of inputs/outputs. SFA, however, often demands larger samples to ensure robust parameter estimation, especially with multiple outputs.

To enhance methodological transparency, we have added a comparative discussion of these models in Section 3.2.1 (SBM-DEA model) and cited key references (e.g., Tone, 2004; Yang et al., 2021) that validate SBM-DEA’s applicability to green efficiency studies.

2. How robust are the results? Did the authors conduct sensitivity analyses to determine whether efficiency scores significantly change under different model specifications?

Response: We sincerely appreciate the reviewer's important question regarding the robustness of our results. To address this concern, we conducted the following sensitivity analyses:

Cross-Efficiency Benchmark Validation. In our prior study (Ma et al., 2021, ), we systematically compared economic efficiency (CTEE), environmental efficiency (CTENE), and green efficiency (CTGE) under identical model frameworks. The results demonstrated that CTGE values exhibited stronger alignment with regional sustainability performance, confirming that green efficiency metrics better capture the multidimensional trade-offs in transportation systems. This cross-metric consistency serves as a critical robustness check for our efficiency measurement framework.

3. How were the input-output indicators selected? Is there empirical justification for including or excluding certain factors?

Response: We appreciate the reviewer’s critical inquiry into the selection of input-output indicators. Our indicator framework was designed based on green development principles and empirical precedents, with rigorous justification for inclusion/exclusion of factors. Below, we clarify the rationale:

Inputs: Transportation infrastructure (e.g., road/rail networks) serves as a necessary but insufficient resource. To generate meaningful outputs, it must combine with complementary factors: Labor (transport sector employees), Capital (fixed asset investments),Energy (fossil fuel/electricity consumption). Supported by Zhou et al. (2018), who emphasize the "infrastructure + enabling factors" framework for green efficiency analysis.

Outputs: Desirable Outputs:Economic: Value-added of transportation (direct GDP contribution). Social: A composite index covering standard of living, Urbanization rate, traffic capacity, and innovation in science and technology. Undesirable Output: Carbon emissions from transport operations. Addresses the limitation of prior studies focusing solely on economic outputs and environmental pollution (Wang et al., 2020).

The selection aligns with the resource-service transformation paradigm in transportation systems. To enhance reader clarity, we have explicitly listed the specific input-output indicators and provided a comprehensive justification for their selection in Section 3.1 (Indicator selection and data source). Please see page 10, line 411-435.

4. Granger causality primarily establishes temporal rather than spatial relationships. Have the authors considered spatial econometric techniques, such as Spatial Autoregressive (SAR) models or Geographically Weighted Regression (GWR), to confirm their findings?

Response: We sincerely appreciate the reviewer’s valuable suggestion regarding spatial econometric techniques. We have supplemented the rationale for methodological selection in Section 3.2.3, providing a detailed justification for the chosen approach in alignment with the research objectives and data characteristics. Below, we clarify the rationale for adopting Granger causality tests over SAR/GWR models in this study:

Firstly, Methodological Alignment with Research Objectives. Our goal was to identify directional causal relationships (e.g., whether Province A’s CTGE improvements systematically precede or influence Province B’s) rather than static spatial spillovers. Granger causality, embedded in the Vector Autoregression (VAR) framework, is uniquely suited for this purpose as it quantifies lead-lag dynamics without requiring pre-specified spatial weight matrices (Huang et al., 2020). In contrast, SAR/GWR models primarily capture contemporaneous spatial correlations (e.g., adjacency effects) but struggle to disentangle temporal causality chains, which are critical for policy sequencing in transportation decarbonization.

Secondly, Limitations of Gravity Models in Efficiency Contexts. While gravity models (distance-based) are widely used to measure spatial linkages (e.g., trade flows), they rely on simplified assumptions (e.g., bilateral distance inversely proportional to interaction intensity) that poorly align with the multi-parametric nature of CTGE. As CTGE is a composite metric derived from input-output optimization, its interregional linkages involve nonlinear, non-distance-dependent mechanisms (e.g., technology diffusion, policy mimicry), which gravity models cannot adequately capture.

Thirdly, Advantages of VAR-Based Granger Causality. The VAR framework allows endogenous determination of interaction directions and magnitudes across multiple regions simultaneously, avoiding the subjectivity of defining spatial weights (e.g., adjacency, inverse distance). By constructing a causality network (rather than a correlation matrix), we derive policy-relevant insights into who drives whom in the CTGE system—key for targeting nodal regions in coordinated governance.

Thank you once again for your invaluable suggestions, which provide significant inspiration for our future research. In subsequent studies, we will rigorously address the methodological limitations by incorporating multiple spatial relationship identification approaches (e.g., gravity models, SAR, and SDM) to conduct comprehensive comparative analyses, thereby identifying diverse spatial correlation network structures of CTGE. Furthermore, we will explore integrating the Spatial Durbin Model (SDM) with Granger causality tests and leverage advanced frameworks such as the Spatiotemporal Panel VAR proposed by Zhou et al. (2019) to disentangle complex spatiotemporal interdependencies.

5.The paper does not explicitly mention whether exogeneity concerns were addressed in the causality tests. Can the authors clarify how they ruled out omitted variable bias or reverse causality?

Response: We sincerely thank the reviewer for raising critical concerns about exogeneity. In response, we have supplemented detailed methodological procedures and results in Section 3.2.3 to clarify how we addressed omitted variable bias and reverse causality through rigorous pre-testing and robustness checks. Below is a summary of the key steps: Firstly, Stationarity and Unit Root Tests. To ensure data stability and avoid spurious regression, we conducted unit root tests using three methods in EViews 8.0: Levin-Lin-Chu (LLC) test (common unit root), Im-Pesaran-Shin (IPS) test (individual unit root), ADF-Fisher test. Results show that the first-difference series of CTGE for all regions passed stationarity tests at the 1% significance level, confirming that the data are I(1) (integrated of order 1) and suitable for further analysis. Secondly, Optimal Lag Selection. The lag order for Granger causality tests was determined by minimizing five criteria: Likelihood Ratio (LR), Final Prediction Error (FPE), Akaike Information Criterion (AIC), Schwarz Criterion (SC), Hannan-Quinn Criterion (HQ). Relationships between regions were retained only if the F-statistic was significant at the 10% level, ensuring robustness against overfitting. Finally, Cointegration Analysis. We employed the Johansen cointegration test to verify long-term equilibrium relationships among regional CTGE values. Trace statistic and max-eigenvalue tests rejected the null hypothesis of no cointegration (p < 0.05), confirming stable interdependencies across regions. This revision strengthens the methodological rigor and aligns with best practices in causal inference for spatial-temporal studies. We deeply appreciate your guidance in enhancing the paper's validity. Please see page 8, line 331-347.

6.What are the economic and policy implications of being a "peripheral" province in the CTGE network? How can western and central regions improve their efficiency?

Response: Thank you for your valuable suggestions. Provinces in the "peripheral" areas of the CTGE network, mainly in the central and western regions, often face challenges in economic development. These regions typically have lower levels of economic activity and development potential compared to the "core" regions in the east. They tend to be more reliant on external support and resources, with limited capacity for independent growth. This is because they mainly receive spill - over effects from other regions rather than actively driving the development of the entire network. From a policy perspective, being a "peripheral" province means that local policies need to focus on improving the region's position in the CTGE network. There is a need for policies that promote investment in transportation infrastructure to enhance connectivity with other regions. Additionally, policies should aim to attract technology and talent from more developed regions. This could involve offering incentives such as tax breaks for businesses that bring in advanced transportation - related technologies or providing housing and education benefits for skilled workers. Moreover, these regions need to collaborate more closely with central government policies to ensure that they can fully utilize national resources and support for development.

In fact, we have already provided some explanations in the Policy implications section, but they are not sufficient. Based on your suggestions, we have further enriched the Policy implications. Please refer to lines 730-801 of the revised draft for details.

7.What role do regional policies, infrastructure investments, and technology diffusion play in shaping spillover effects? Can the authors provide empirical examples of successful interventions?

Response: We are sincerely indebted to the reviewer for their thorough and meticulous work. In our paper, during the analysis of Degree Centrality, we posited that "in economically advanced regions, the transportation network is well - developed and the technological level is relatively high. As a result, production factors like capital and technology can more readily spill over to regions with lower comprehensive transportation green efficiency." Nevertheless, we failed to elaborate on the specific roles these factors play in generating spillover effects, which has led to some confusion among readers.

In light of your valuable suggestion, we have conducted an in - depth explanation within the paper. We have also referenced relevant academic literature to substantiate our viewpoints. For specific details, kindly consult line 613-625 of the revised manuscript.

8.The authors identify four network blocks (main inflow, main outflow, bidirectional spillover, and agent block), but what specific policy recommendations apply to each block?

Response: We are deeply grateful for your insightful suggestions. In light of these recommendations, we have meticulously revised the policy suggestions section. Specifically, we have formulated targeted countermeasures for different blocks, aiming to enhance the practicality and effectiveness of our proposed strategies. For detailed information, please refer to line 730-801 of the revised manuscript. This revision not only reflects our commitment to addressing the issues raised but also demonstrates our dedication to improving the quality and comprehensiveness of the research. We believe that these refined policy suggestions will contribute more significantly to the field of study and relevant decision - making processes.

9.Why were these specific years chosen? Would including more recent years (e.g., post-pandemic data from 2021–2023) significantly alter the findings?

Response: We are truly grateful for your inquiry. The research period of this paper spans from 2003 to 2020, a choice made with several crucial factors in mind. Primarily, by setting the end - point at the conclusion of the "13th Five - Year Plan" (in 2020), it becomes feasible to conduct a comprehensive assessment of the efficacy of the "Transportation Power" strategy. This approach enables us to evaluate the long - term impacts and achievements of the strategic initiatives implemented during this period.

Furthermore, it is essential to note that we deliberately excluded the data from 2021 to 2023 to avoid the distorting effects of the pandemic. The data during these years exhibited significant and abnormal fluctuations. For example, in the first quarter of 2022, the consumption of diesel declined by 22% year - on - year. Such fluctuations could potentially obscure the underlying long - term trends that are central to our research objectives.

Regarding data currency, we fully recognize its importance. To this end, we are actively engaged in the collection and collation of data for the years 2021 - 2023. Once the data has been thoroughly processed and analyzed, we will promptly release our updated research findings. This will ensure that our research remains relevant and reflects the most current state of the transportation industry.

We have also taken the initiative to address this limitation in Section 5.3 of the paper. By doing so, we aim to provide transparency and clarity to our readers, allowing them to better understand the scope and potential limitations of our study. Once again, we express our sincere appreciation for your valuable input.

10.How reliable is the transportation industry data? Was any data imputation performed for missing values?

Response: Thank you very much for your valuable suggestions. They are somewhat similar to those of the first reviewer. To ensure the authority and accessibility of the data, all our data are sourced from the National Bureau of Statistics of China, including the China Statistical Yearbook, China Energy Statistical Yearbook, and Transport Industry Statistical Bulletin, etc. We obtained most of the data from these sources. For the few missing data in individual regions, we supplemented them using the interpolation method. In response to the suggestions of the two reviewers, we added relevant explanations in the Section 3.3 Indicator selection and data source, we have supplemented the data sources used in this study and included a table presenting the descriptive statistics of the relevant data in the Indicator selection and Data Sources to enhance the readability of the article. We sincerely appreciate your guidance once again. Please see page 10, line 410-433.

11.Are there sectoral variations in efficiency? Would a disaggregated analysis (e.g., road vs. rail vs. aviation) provide deeper insights?

Response: We sincerely appreciate your valuable suggestions. This paper does not conduct measurements on the efficiency of each sector within the transportation industry, primarily for the following reasons.

Firstly, the focus of this research is on the green efficiency of China's comprehensive transportation. It encompasses statistical data from diverse transportation modalities, such as road, railway, waterway, and aviation. The overarching aim is to uncover the holistic state of the transportation industry across 30 regions in China from a macro - perspective. This aligns seamlessly with the policy - oriented objectives of the "National Comprehensive Three - Dimensional Transportation Network Planning Outline". By taking this comprehensive approach, we can better understand the overall performance and development trends of the transportation industry at a national scale, which is crucial for formulating effective industry - wide policies and strategies.

Secondly, the inconsistent statistical granularity poses a significant obstacle to obtaining sector - specific transportation data. For instance, provincial - level energy data are only comprehensively available for roads and railways. In contrast, data for aviation, where airport - level information is not disaggregated by province, and waterway transportation, with its fragmented port - based data, are severely deficient. This lack of consistent and complete data at the sector level makes it extremely challenging to conduct in - depth efficiency measurements for individual transportation sectors.

Nonetheless, your suggestions have been highly inspiring and will play a crucial guiding role in our future research endeavors. We believe that leveraging transportation Internet - of - Things data and applying the network DEA method to allocate efficiency holds great promise. We have elaborated on this aspect in the Limitations Section 5.3 of the paper. This limitation not only acknowledges the current constraints in our research but also points towards potential research directions for future studies, aiming to overcome these challenges and further deepen our understanding of the transportation industry's efficiency at a more detailed level.

12.Literature Review: The paper discusses existing research but would benefit from a more structured comparison of prior studies, particularly those that apply network analysis to transportation efficiency. It is recommended to include additional review papers:

Guo, J., Bai, S., Li, X., Xian, K., Liu, E., Ding, W., & Ma, X. (2025). A universal geography neural network for mobility flow prediction in planning scenarios. Computer‐Aided Civil and Infrastructure Engineering.

Li, A., Xu, Z., Li, W., Chen, Y., & Pan, Y. (2025). Urban Signalized Intersection Traffic State Prediction: A Spatial–Temporal Graph Model Integrating the Cell Transmission Model and Transformer. Applied Sciences, 15(5), 2377.

Response: We are appreciated for your professional comments. After reviewing the literature, we cited more recent and pertinent research. For example:

Eun Hak Lee. eXplainable DEA approach for evaluating performance of public transport origin-destination pairs[J]. Research in Transportation Economics, 2024, 108: 101491. https://doi.org/10.1016/j.retrec.2024.101491

Guo J, Nakamura F, Li Q, et al. Efficiency Assessment of Transit-Oriented Development by Data Envelopment Analysis: Case Study on the Den-en Toshi Line in Japan [J]. Journal of Advanced Transportation, 2018, 6701484. https://doi.org/10.1155/2018/6701484.

Guo, J., Bai, S., Li, X., Xian, K., Liu, E., Ding, W., & Ma, X. A universal geography neural network for mobility flow prediction in planning scenarios. Computer‐Aided Civil and Infrastructure Engineering. 2025, 13398. https://doi.org/10.1111/mice.13398.

Li, A., Xu, Z., Li, W., Chen, Y., & Pan, Y. . Urban Signalized Intersection Traffic State Prediction: A Spatial–Temporal Graph Model Integrating the Cell Transmission Model and Transformer. Applied Sciences, 2025, 15(5), 2377. https://doi.org/10.3390/app15052377.

Li, S J., Ma, Q F. Carbon quota allocation and emission reduction responsibility sharing at provincial level in China from transport industry [J]. Sustainable Futures, 2025, 9: 100535. https://doi.org/10.1016/j.sftr.2025.100535.

Ma, H. T. Urbanization under globalization: How does the Belt and Road Initiative affect urbanization levels in participating countries [J]. Journal of Geographical Sciences, 2022, 32 (11): 2170-2188. https://doi.org/10.1007/s11442-022-2042-1.

Cao, Y. W., Zhang, R. R., Zhang, D. H., et al. Urban Agglomerations in China: Characteristics and Influencing Factors of Population Agglomeration[J]. Chinese Geographical Science, 2023, 33(04): 719-735. https://doi.org/10.1007/s11769-023-1368-7.

Upadhyay, R. K., Sharma, S. K., Kumar, V. Introduction to Intelligent Transportation System and Advanced Technology[J]. Energy, Environment, and Sustainability, Springer, Singapore, 2024. https://doi.org/10.1007/978-981-97-0515-3_1.

13. Please check the name of "Shaanxi" in Figure 4.

Response: Thank you for your feedback. The official English name of 陕西 is Shaanxi, rather than the Hanyu Pinyin form Shanxi. This distinction was established primarily to differentiate it from another province that shares the same Pinyin transliteration "Shanxi" (山西, Shanxi Province). In China, most provincial English names follow Hanyu Pinyin conventions. However, since English lacks tonal markers, an extra letter "a" was intentionally added to the English name of 陕西 (Shaanxi) to avoid confusion with 山西 (Shanxi). This naming convention aligns with international standards for Chinese geographical name standardization (GB/T 17693.1-2022) and has been officially endorsed by the Chinese government and the United Nations Group of Experts on Geographical Names (UNGEGN).

14. English expressions do not use full stops in certain contexts, so the symbol in Figure 6 should be a comma instead of a Chinese character.

Response: Thank you very much for your meticulous and thoughtful feedback, which has been invaluable in enhancing the quality of our manuscript. In accordance with your recommendation, we have revised the erroneous punctuation marks in Figure 7 (originally labeled as Figure 6 in the previous version), replacing Chinese enumeration commas (、) with standard commas (，). We sincerely appreciate your expert guidance once again.

15. Please check the entire manuscript for grammatical accuracy.

Response: Thankful to your careful reading and valuable suggestion. We have invited the native English speaker of this field to improve the language of this manuscript. Thank again for your valuable suggestion.

 

Finally, we would like to express our great appreciation and respect to you. According to your professional and helpful comments, our paper has made great improvement. We have carefully revised and responded to each comment raised by reviewer. Finally, we show great respect for reviewer’s professionalism. If we have any unclear explanations, please contact us immediately. Thank again for your positive and useful comments.

Yours Sincerely

The authors

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I am happy with the responses.

Reviewer 2 Report

Comments and Suggestions for Authors All my concerns have been addressed well.