Assessing Mobility-Driven Socio-Economic Impacts on Quality of Life in Small Urban Areas: A Case Study of the Great Belt Fixed Link Corridor

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

Thank you for the opportunity to review the paper. This is an excellent paper with high publication potential. The paper effectively combines theories of time-space compression, mobility justice, and the mobility paradox, situating findings within broader urban theoretical discourses.

This paper presents a robust and well-documented investigation into the long-term socio-economic impacts of the Great Belt Fixed Link on small urban areas in Denmark. The paper is commendable for its integration of spatial analysis, socioeconomic indicators, and predictive modelling using ANN-based land use forecasting. It stands as a significant contribution to both the transport geography and urban planning literature, particularly in the domain of fixed-link infrastructures and their effects on regional cohesion and spatial equity.

The combination of survey data, spatial GIS analytics, and machine learning predictive models is methodologically strong and exemplary in terms of interdisciplinary integration. The paper uses extensive, high-quality datasets (e.g., CORINE, national statistics), which lend strong empirical support to its findings.

Although I evaluate the paper positively, I suggest you consider the following points: 

• The paper refers to net population change without addressing intra-regional or inter-municipal migration flows, limiting insights into demographic dynamics and residential mobility. I suggest incorporating microdata or register-based migration flows (if possible or in future research) to strengthen conclusions on population redistribution.
• While the survey provides valuable attitudinal data, the paper lacks qualitative depth (e.g., interviews, case narratives) to unpack the lived experiences and perceptions of exclusion (e.g., toll burdens). Future research could adopt ethnographic or longitudinal interviews to explore behavioural adaptation to mobility infrastructure.
• Some sections, especially the introduction and literature review, contain repetitive phrasing (e.g., "mobility connectivity", "socio-economic dynamics"). I suggest streamlining terminology for clarity and precision (e.g., use “fixed-link effects on spatial equity” instead of longer paraphrases).
• I suggest improving figure clarity (enlarge and expand them to the entire page width). Figures (e.g. Figures 2 and 3) are not always clearly labelled or referenced consistently in the text. Multiple sub-figures are marked as a, b, c, and d, creating potential confusion.

Author Response

We thank the reviewer for his/her constructive comments and suggestions, which have significantly helped us improve the manuscript. Below, we provide detailed responses to each question and point raised, indicating how we have revised the manuscript text or provided explanations for specific comments to the reviewers. All modifications are highlighted in the revised version of the manuscript.

_______________________________

I suggest you consider the following points: 

• The paper refers to net population change without addressing intra-regional or inter-municipal migration flows, limiting insights into demographic dynamics and residential mobility. I suggest incorporating microdata or register-based migration flows (if possible or in future research) to strengthen conclusions on population redistribution.

Thank you for this important observation; we agree with your comment. We acknowledge that our current analysis is limited to aggregate population change and does not capture the complexity of intra-regional migration. Although detailed register-based migration data were not accessible during this study, we now explicitly highlight this limitation in Section 5.4.

 

• While the survey provides valuable attitudinal data, the paper lacks qualitative depth (e.g., interviews, case narratives) to unpack the lived experiences and perceptions of exclusion (e.g., toll burdens). Future research could adopt ethnographic or longitudinal interviews to explore behavioural adaptation to mobility infrastructure.

We appreciate this insightful comment. We agree that integrating qualitative depth would enhance the understanding of behavioural adaptation and perceived exclusion, particularly regarding toll burdens and spatial inequity. However, due to constraints of scope and time resources, this study focused on spatial modelling and survey-based data. In response to this point, we have now revised section 5.4.

 

• Some sections, especially the introduction and literature review, contain repetitive phrasing (e.g., "mobility connectivity", "socio-economic dynamics"). I suggest streamlining terminology for clarity and precision (e.g., use “fixed-link effects on spatial equity” instead of longer paraphrases).

We followed the reviewers' recommendations, we have modified certain phrases within the text.

The term ‘socio-economic dynamics’ has been mentioned solely in the abstract and the introduction. The instance in the introduction has now been amended to ‘socio-economic impact.’

We do not use the phrase ‘mobility connectivity’ together in the manuscript; however, they are employed individually as they represent keywords within this manuscript.

 

• I suggest improving figure clarity (enlarge and expand them to the entire page width). Figures (e.g. Figures 2 and 3) are not always clearly labelled or referenced consistently in the text. Multiple sub-figures are marked as a, b, c, and d, creating potential confusion.

All figures in the paper have been replaced with newer, higher-resolution images. The initial arrangement labelled as ‘a, b, c, and d’ has been removed as it is no longer required. The figures have been enlarged to fit the page width. Furthermore, as this publication is in digital format, these high-quality images will enable readers to zoom in and see even more details in the images.

Reviewer 2 Report

Comments and Suggestions for Authors

Changes need to be made

Comments for author File: Comments.pdf

Author Response

We thank the reviewer for his/her constructive comments and suggestions, which have significantly helped us improve the manuscript. Below, we provide detailed responses to each question and point raised, indicating how we have revised the manuscript text or provided explanations for specific comments to the reviewers. All modifications are highlighted in the revised version of the manuscript.

___________________________________

Introduction

The structure of the introduction is consistent, and the transition from the task statement to the literature review is logical. - The introductory part sets the framework, but it is worth once again clearly defining the statement of the scientific task at the end of the first paragraph — now it is "lost" between the enumeration of approaches.

We have now revised the final paragraph of the introduction to repeat the focus of the manuscript.  -  ‘…by evaluating how the Great Belt corridor, as a fixed link, has influenced urban development and socio-economic impact, particularly with regard to spatial equity, land use, housing, and mobility transformations across regional and local scales within the municipalities situated on either side…’

 

- Descriptions of previously performed works are heavily overloaded with enumerations of aspects — it is worth structuring them by subsectors: 1) impact on land use, 2) impact on mobility, 3) on the housing market, 4) on the standard of living.

We have restructured the literature review section and divided it into two parts.

- It would be nice to give short interim conclusions (based on the results of each paragraph) so that the reader understands why this is important for the current study. Scientific novelty and personal contribution

We appreciate the reviewer’s suggestion regarding the inclusion of interim conclusions. In response, we've added concise summary paragraphs at the end of each subsection in the literature review. These summaries highlight the key insights from the reviewed studies and clarify how they contribute to the current research. Specifically, they articulate the scientific relevance and novelty of our approach, which combines predictive modelling and socio-demographic analysis to assess the long-term impacts of GBFL.

 

- It is important for the introduction and review of the literature to highlight your contribution in comparison with the already known one. The current version does not sufficiently emphasize how the new study differs from previous ones (for example, "most studies focus on macroeconomic effects, while we focus on small towns; or, for the first time, the analysis of the housing market, migration, and urban morphology is combined").

Thank you for your remark. While we have already included this segment of our contribution in the introduction, it is highlighted even more in this revised version. However, considering the suggestions from the other reviewers, we prefer not to be overly categorical by stating ‘for the first time….’

However, we would like to emphasise that this study shifts the focus towards small and medium-sized municipalities located on both sides of the GBFL, examining how long-term infrastructural investments reshape social and spatial conditions at a more detailed territorial scale. To the best of our memory, no other studies have the same design and methodological setup.

 

- It would be good to highlight not only what is being done, but also "why the previous approaches are not enough."

While previous assessments of the GBFL have addressed specific areas such as traffic volumes, engineering feasibility, or macroeconomic cost-benefit analyses, they often lacked a spatially explicit and socio-economically integrated perspective. In particular, they did not systematically analyse how the GBFL reconfigures regional accessibility patterns, urban land use, or local development conditions over time.

We have this stetememnt in the introduction and ate the nd of the literature review section:

This study bridges that gap by employing spatial predictive modelling (ANN–CA) alongside socio-economic datasets and spatial analysis. This integrated approach enables us to simulate and map not only physical land use changes but also to trace behavioural adaptations (e.g., commuting or mobility patterns) and the dynamics of territorial cohesion. Such a methodology has not been previously applied to fixed links in Denmark, making our contribution unique.

 

Methods

Each paragraph should focus on one idea: either a description of the method, a description of the data, or a description of the processing mode.

We have now expanded the methodology section and compiled all the methods utilised therein. The collected results are presented in the results section, while the limitations and disadvantages associated with the study have been relocated to the discussion.

 

- Explain the tools and steps: “isochron analysis", “network analysis” — which programs, which software (GIS? QGIS? ArcGIS?), what is the approach, time slice, detail of spatial cells.

This is now clearly noted and explained in the methodology section:

‘Isochrone modelling for commuting zones was conducted using the OSMnx and NetworkX Python libraries within QGIS and Jupyter Notebook environments. Commuting zones were generated using an OpenStreetMap-derived road network…’

 

- List all the indicators in one list, explain why they are selected.

We have listed all the indicators in Table 1.

 

- Emphasize the advantages and limitations of the data used.

In this regard, we are following the 4^th reviewer's comments and are addressing all limitations, including data, in the discussion section.

 

- If you mention “scenario analysis”, “retrospective assessment”, explain what you mean: individual models? calibration based on history and application to the forecast?

Upon reviewing the document, we were unable to locate the term - ‘scenario analysis,’ despite our use of scenario simulation or forecasting, which is evidently related to the model.

 

The term ‘retrospective assessment’ - refers to the analysis of historical trends, serving merely as a synonym for the term 'historical.'  The sentence in the literature review has now been changed to: ‘…retrospective assessments (historical land use and socio-economic trends) and predictive...’

 

 

Results

Logical connections. Several hypotheses have been proposed (the impact of the pandemic, gender differences, and the work–life balance effect). It is advisable to back them up with literature or note that this is an assumption that requires further research.

We agree with the reviewer, we have now added a new sentence in the results section:

‘Although observed trends support such interpretations, they remain hypothetical and require further empirical validation through longitudinal or behavioural studies.’

 

- Causal conclusions. In some places, the conclusions are extremely categorical ("improved access not only reduces time... but it also provides significant economic benefits ..."). Excessive categoricality should be avoided if the conclusions are not supported by calculations/statistics.

We agree with the reviewer. However, we have now included the reference for this section, in which we reference the Danish Ministry of Transport and the road directorate from where the knowledge has been derived over the years based on the official reports.

 

- The context. It is correctly noted about the need to manage negative consequences (traffic jams, accidents), but this aspect requires at least a short reference to research or data. 5. Details and specifics

Thanks to the reviewer for the good observation, we have now added references and one extra sentence to the section, saying:

‘This aligns with other studies showing that higher congestion levels correlate with increased total and serious injury crashes, indicating the need for effective safety management strategies’

 

- Provide information on the survey. If possible, indicate how many respondents were interviewed — this is important for assessing the validity of the percentages given.

The survey we conducted served primarily to validate information that was already known in relation to the official governmental surveys that preceded our efforts. Given that it does not offer any new insights or scientific value (in terms of saying something new rather than what we already know) for this research, we have chosen to concentrate on other, more significant aspects of the study. We will refer to this data, but will not be looking into creating graphs and describing them.  

 

- More explicit links to previous sections. Note that the isochronous zone analysis and the survey result are integrated. 6. Errors and typos - "increased risk of road accidents" — if possible, indicate whether there are grounds for this in your data or this is a fact from external research.

This statement is backed by research from the Danish Ministry of Transport. We’ve also included a reference to support our claim.

 

- "potential reduction in CO2 emissions" — it is worth clarifying whether this is calculated by you, or is it a theoretical effect noted in the literature.

There are no direct calculations or modelling of CO₂ emissions conducted by us within this study, considering the extent or complexity of this research could have reached. However, environmental impacts and climate concerns are discussed theoretically in relation to infrastructure planning by official governmental bodies due to climate change-related issues. And GBFL is not the exception.

 

 

Discussion

It is important to clearly indicate where empirical data comes from and how they compare with the theoretical provisions and statistical results presented.

We appreciate the reviewer’s valuable suggestion. The empirical data used in this study are clearly outlined in Table 1 of the manuscript, which details the sources of the empirical data.

These empirical inputs are often compared against relevant theoretical frameworks, for instance, in section 5.2. ‘…accessibility does not automatically result in equitable development unless supported by proactive planning and institutional coordination [52], which is consistent with findings in [53] and [54]. The primary…’

 

- There are many statements that need either concrete confirmation (for example, about the differentiation of growth in Odense, Ringsted, Slagels), or references to other works.

If I correctly understand, the reviewer refers to section 5.2, where we say: ‘….the majority of municipalities analysed showed a reduction in unemployment rates and some degree of population growth, notably in the municipalities of Odense, Ringsted, and Slagelse…’ We have now revised this part and provided references.

 

- Terminology not previously disclosed is sometimes used ("opportunity structures", "cumulative benefits", etc.). 4. The problem of causality and accuracy of conclusions

Thank you for your thorough observation – ‘opportunity structure’ was replaced by ‘opportunities’, and ‘cumulative benefits’ – has been removed from the manuscript during the modifications.

 

- A number of interpretations are too categorical ("These changes were not accompanied by a uniform increase in housing prices...", "population growth increased"). The statistical significance should be indicated (if calculated), or the conclusion should be presented in a more moderate form ("it was noticed", "a trend towards ..." was detected).

We have reviewed the document and modified it with proposed interpretations. We also employed the term ‘it was noticed’ in the conclusion section to make it less categorical. In several instances, we also included official references to support the information.  

 

- The mechanism why infrastructural improvement does not lead to the desired equity is not disclosed — we can add a brief analysis of the structural causes and the impact of other factors (housing availability, market constraints, differences in institutional support, etc.). 5. Scientific novelty and contribution of the section

We agree that factors such as housing availability, institutional support, and market constraints play critical roles in shaping the socio-economic effects of large-scale infrastructure. However, a comprehensive structural analysis of these dimensions would necessitate a different methodological design, including qualitative fieldwork and more detailed housing market data that extend beyond the scope of our current spatial and predictive modelling framework.

We have therefore acknowledged this as a limitation in the revised manuscript (Section 5.4 – ‘…although the study identifies persistent spatial disparities despite enhanced connectivity, it lacks a comprehensive analysis of the underlying structural mechanisms, such as housing…’) and recommended it as a key direction for future research. This clarifies the boundaries of our study while also recognising the value of such an extension in deepening the understanding of equity-oriented planning within the context of infrastructural development.

  

- It is good that the paradox of mobility and the need for long-term strategic planning are mentioned. It would be even better if you correlated your results more explicitly with existing research (what is confirmed, what contradicts, what adds?).

Thank you for this valuable suggestion. In response, we have now enhanced the text in section 5.2 to respond on the reviewer’s comments.

‘…These findings are generally consistent with previous studies [51], [52], [53], [54], [55], highlighting that while transport infrastructure improves accessibility, it does not inherently guarantee equitable socio-economic outcomes unless supported by long-term strategic planning. The results of our study indicate that socio-economic benefits remain spatially concentrated, particularly in municipalities that are already advantaged. Consequently, this research contributes to the discourse by integrating predictive modelling, land use change, and commuting behaviour to demonstrate that, despite the decentralised relocation of jobs and enhanced commuting potential, regional disparities in housing affordability and population growth persist. These findings underscore the significance of institutional mechanisms and cross-sectoral planning to prevent the reinforcement of existing regional imbalances.’

 

- If possible, emphasize the methodological limitations of your analysis: what data was sufficient for, which requires further research (for example, qualitative longitudinal observations, study of small settlements, etc.). 6. Technical and design aspects

Thank you for noticing this gap. We have now incorporated a ‘5.4 Methodological Considerations and Future Research Directions’ section in the discussion, addressing the data's limitations as well as future research directions.

 

- The figures ("transfer of approximately 5,750 public sector facilities") need to be clarified — are we talking about jobs or physically displaced institutions?

‘5,750 individual state jobs...’ this has now been corrected in the manuscript.

- It is better to replace phrases like "there is a decrease in the level of nonemployment" with "there is a decrease in the unemployment rate." Conclusions

The manuscript no longer contains any ‘nonemplyment’ word.

- There is not enough evidence to explain why urban growth is considered modest but sustainable: we need clear numerical, cartographic or comparative indicators, links to your data or literature.

Thank you for this observation. We have clarified the basis for characterising urban growth as “modest but sustainable” by explicitly referencing both quantitative indicators and cartographic simulations derived from our land use modelling results. While this classification may differ in comparison to other national contexts, our assessment is grounded in the Danish demographic, and planning framework, which prioritises controlled and incremental urban development.

 

- It is not disclosed how transport corridors contribute to this type of growth — it is advisable to rely on a study of the dynamics of accessibility and migration.

We have consistently cited the research carried out by the transport ministry and official governmental bodies that regularly undertake studies and publish reports on various aspects across the country, including accessibility and migration.

 

- A comparison with other regions or periods would make the conclusion more significant.

We did make a few comparisons with other studies throughout the manuscript to describe the data derived from the model; however, we also address this in the conclusion.

 

- Specification of the results is required: provide metrics (accuracy, Kappa coefficient and an explanation of what it is for the reader).

We have specifications in the methodology, and we again repeat them through the manuscript in the relevant places as a reminder to the reader.

 

- It is necessary to explain why ANN was chosen and how it is superior to alternatives in the context of your spatial tasks.

This detail is now indicated in the methodology section.

 

- Emphasize that the model reflects only current (existing) conditions and does not take into account possible future changes in conditions. The conclusions correctly emphasize the importance of infrastructure, but do not reveal what other socio-economic factors influence limited growth (e.g. demography, politics, economics, migration).

We did mention this in the conclusion; however, we could clarify it further by revising the paragraph.

 

- For a stronger justification, add a brief list of specific barriers: lack of demand for relocation, limited housing construction, etc.

This has been included in the conclusion section.

 

- Make it clear that transport projects are only part of a complex set of development factors.

We have again included a statement about it in the conclusion.

- The recommendations remain abstract (“context-sensitive pricing models", “proactive planning"). Specify: which tariffs, for whom, and how to integrate predictive models into real solutions?

While we recognise the importance of detailed socio-economic recommendations particularly concerning toll pricing and accessibility, our study did not include the level of qualitative data (e.g., interviews, income-specific analysis) necessary to support concrete, group-specific policy prescriptions. Therefore, we deliberately refrain from issuing categorical statements in this regard. Nevertheless, it is worth noting that discussions on toll adjustments and equity-sensitive pricing are already underway in public and policy forums, suggesting that our findings can contribute to these broader debates.

- Show how your findings can be used by municipalities/the state for practical planning and management. - It is not sufficiently detailed how (in which territories, for which groups) the uneven benefits of infrastructure manifest themselves.

We have now included this aspect in the 5.3 section.

 

- It should be clearly shown how these differences can be mitigated (for example, special support programs, investments in “losing” areas). Limitations of the study

We have now added the text in the 5.4 section, addressing this aspect.

- It is good that the limitations are given (geography, aggregated data, the inability to fully detail migration flows), but describe how this may affect the reliability and interpretation of the conclusions.

We have now added the text in section 5.4 to address this remark.

- Describe what future research can remove these restrictions (for example, the use of mobile data, more detailed surveys, inter-regional comparison).

 

We have now included the text in section 5.4 to address the reviewers' remarks.

 

 

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript evaluates spatial and socioeconomic consequences of linking with a fixed transport system two localities in Denmark. The authors argue the need to clarify a gap in how transport corridors modify the shape of connectivity between urban regions. They analyze the socioeconomic effects of the modified connectivity. The paper uses spatial modeling to investigate commuting habits in a time frame of 28 years within a 45-minute travel time radius.

Can you provide more details about the ANN prediction model? What type of layers were used? How was the data structured for training?

Can you describe in detail the cellular automata simulation?

In Section 4.4, can you explain in detail why correlation was used to eliminate features? What is the theoretical basis for this? Since this involves a form of prediction, the removed features might have contributed to understanding how things change over time. The presence of correlation could have helped the network make more confident predictions.

In Section 4.4, what is meant by "standard learning parameters"? What standard was followed? What parameter values were used?

It’s not entirely clear—does the text state that the model was trained with data from 2012 to 2018 and then used for predictions from 2024 onward? Why wasn’t data up to 2024 used?

In Figure 4, the images are too small, and all the figures appear nearly identical. Additionally, there’s a labeling error—all images are marked as (a) and (b).

It’s quite noticeable that the prediction shows almost no expansion. In predictive modeling, it’s uncommon for models to achieve "stabilization" as the authors claim. Instead, it seems the model failed to make meaningful predictions and simply replicated the input data with minimal changes, which is why the figures in Figure 4 look so similar. Please provide more detail on this aspect.

The authors should include images from previous years (e.g., 2010–2018) and compare them with the predicted series (2024–2036). This would show whether the trend remains consistent or, if not, allow for a detailed explanation of why the model does not follow the same trend.

In line 202, use a capital letter to start the title 

Finally, what is the meaning of GBFL on the manuscript? The Great Belt Fixed Link or the Great British Fixed Link? According to the manuscript, is the latter is it correct? Please clarify.

Please state if you have used IA applications like ChatGPT or Deep Seek to write this manuscript, and to what extent.

Comments on the Quality of English Language

Please state if you have used IA applications like ChatGPT or Deep Seek to write this manuscript, and to what extent.

Author Response

We thank the reviewer for his/her constructive comments and suggestions, which have significantly helped us improve the manuscript. Below, we provide detailed responses to each question and point raised, indicating how we have revised the manuscript text or provided explanations for specific comments to the reviewers. All modifications are highlighted in the revised version of the manuscript.

______________________________________

 

Comments and Suggestions for Authors

The manuscript evaluates spatial and socioeconomic consequences of linking with a fixed transport system two localities in Denmark. The authors argue the need to clarify a gap in how transport corridors modify the shape of connectivity between urban regions. They analyze the socioeconomic effects of the modified connectivity. The paper uses spatial modeling to investigate commuting habits in a time frame of 28 years within a 45-minute travel time radius.

Can you provide more details about the ANN prediction model? What type of layers were used? How was the data structured for training? Can you describe in detail the cellular automata simulation?

We have now rewritten the entire 3.3 section of the methodology and provided a detailed description of the ANN prediction model, including its layers, data structure, and training, while also elaborating on the cellular automata simulation.

 

In Section 4.4, can you explain in detail why correlation was used to eliminate features?

We practically stated that feature elimination was done to prevent multicollinearity. In the text, in the methodology section that is now very detailed and describes everything in detail.

 

 

What is the theoretical basis for this?

The theoretical reasoning is actually tied to ANN learning dynamics, generalisation, and convergence. We have now described this in the methodology section.

 

 

Since this involves a form of prediction, the removed features might have contributed to understanding how things change over time.

While removed features (like housing prices) could carry value, they were excluded due to redundancy and risk of noise, not simply correlation. We have addressed this in the methodology section as well.

 

 

The presence of correlation could have helped the network make more confident predictions.

In the methodology section, we now explain that keeping highly correlated variables can hurt model performance rather than enhance it.

 

 

In Section 4.4, what is meant by "standard learning parameters"? What standard was followed? What parameter values were used?

We appreciate the reviewer’s thorough feedback and acknowledge that ’standard learning parameters’, which refer to the training, were not sufficiently specific in our manuscript. We have now clarified those parameters in the methodology section.  

 

 

It’s not entirely clear—does the text state that the model was trained with data from 2012 to 2018 and then used for predictions from 2024 onward? Why wasn’t data up to 2024 used?

The model was indeed trained on land use transitions observed between 2012 and 2018 and subsequently used to simulate future scenarios for 2024, 2030, and 2036. The decision not to include data up to 2024 was due to a lack of available and validated CORINE land cover data beyond 2018 at the time of modelling for this research. CORINE provides harmonised EU land use datasets, which are typically updated at intervals of a few years. At the time of this study, the most recent fully processed and quality-assured CORINE dataset was from 2018. Using incomplete or unvalidated datasets beyond this point would have compromised the spatial consistency and comparability required for training and prediction. To make a clear statement about it, in the data collection section 3.2, we now added the sentence to clarify this point:

‘As the most recent validated CORINE dataset available at the time of modelling was from 2018, spatial inputs for predictive modelling were harmonized to this temporal point to maintain consistency and ensure forecasting reliability (Table 1).’

 

 

In Figure 4, the images are too small, and all the figures appear nearly identical. Additionally, there’s a labeling error—all images are marked as (a) and (b).

All figures and images have now been revised; their clarity has been improved, making them more readable. The labels “ a, b, c, and d “ have been removed to avoid any confusion for the reader. Instead, we now utilise the titles on the maps that indicate the year, reflecting the generated map content.

 

 

It’s quite noticeable that the prediction shows almost no expansion. In predictive modelling, it’s uncommon for models to achieve "stabilization" as the authors claim. Instead, it seems the model failed to make meaningful predictions and simply replicated the input data with minimal changes, which is why the figures in Figure 4 look so similar. Please provide more detail on this aspect.

We acknowledge that the predicted maps for 2024, 2030, and 2036 appear visually similar; however, this outcome reflects a combination of factors rather than a failure in prediction. First, the CORINE dataset used for training (2012–2018) showed only minor land use transitions in the study area. As such, the ANN-CA model, trained on real-world low-variation data, reproduced these gradual patterns with high fidelity. Second, as we have already stated in the manuscript, the land use dynamics in Denmark are influenced by regulatory and environmental constraints, leading to slow and highly controlled urban expansion. Third, the CA algorithm used in this research was designed to simulate land conversion only when input thresholds are exceeded, and in a context of low transition intensity, successive iterations result in diminished changes over time.

These numbers and statements have now been clearly clarified in the manuscript.

 

 

The authors should include images from previous years (e.g., 2010–2018) and compare them with the predicted series (2024–2036). This would show whether the trend remains consistent or, if not, allow for a detailed explanation of why the model does not follow the same trend.

In Figure 2 of the urban growth maps, urban expansion between 1990 and 2018 is clearly evident. Figure 4 presents the maps from 2010 and 2024, illustrating the observable changes leading up to 2024. It is noteworthy that there was not a significant difference between 2018 and 2024; therefore, we preferred to utilise the model from 2024. In the results section, we now provide a more detailed examination of the parameters of land use change between 2010 and 2024.

 

 

In line 202, use a capital letter to start the title 

The letter case in the title has now been corrected.

 

 

Finally, what is the meaning of GBFL on the manuscript? The Great Belt Fixed Link or the Great British Fixed Link? According to the manuscript, is the latter is it correct? Please clarify.

Thank you for your remark; we have now corrected the typographical error. We have defined the Great Belt Fixed Link as GBFL in the text, and we are referring to it throughout the document.

 

 

Please state if you have used IA applications like ChatGPT or Deep Seek to write this manuscript, and to what extent.

We have now made a statement at the end of the document regarding the use of AI.

 

Reviewer 4 Report

Comments and Suggestions for Authors

Review of "Assessing Mobility-Driven Socio-Economic Impacts on Quality of Life in Small Urban Areas: A Case Study of the Great Belt Fixed Link Corridor"

Thank you for allowing me to review your paper. It is an interesting and valid study, evaluatingthe impacts of the Great Belt Fixed Link (GBFL) on smaller urban areas through spatial modeling, surveys, and socio-economic analysis. The study’s findings are of interest because they highlight the GBFL’s role in revealing uneven socio-economic benefits. That being said, I have concerns with this study:

1. Isochrone Justification:
   The study’s scope is defined by a 45-minute commuting isochrone, but there is no justification for this choice, nor are sensitivity tests conducted with alternative isochrones (e.g., 30 or 60 minutes). This selection undermines the robustness as different isochrones could alter commuting patterns, urban growth estimates, and socio-economic impacts, limiting confidence in the results’ generalizability.
   
   
2. Missing ANN Model Parameters:
   The Artificial Neural Network model’s input parameters (e.g., number of hidden layers, learning rate, activation functions) are not provided, despite the model being central to the predictive analysis. These parameters are crucial for ensuring reproducibility.
   
   
3. Lack of Formal Hypotheses:
   The study poses broad research questions but lacks a formal hypothesis to guide the analysis. This makes the study appear descriptive rather than hypothesis-driven. It makes it harder to evaluate findings against specific expectations.
   
   
4. Data Preprocessing and Temporal Misalignment:
   The data collection description lists sources but omits critical details on data preprocessing steps, and integration methods for geospatial and survey data. Additionally, using 2024 data for a 1990–2018 study introduces temporal misalignment without justification. One must infer the reason for the dates selected.
   
   
5. Superficial Analysis of Drivers and Policy Implications:
   The study identifies uneven socio-economic benefits across municipalities but provides a superficial analysis of the drivers behind these disparities (e.g., local economic conditions, toll costs) and fails to propose targeted policy interventions. Greater insights into the causes of outcomes and specific policies (e.g., subsidies for lagging regions) would enhance the study’s practical relevance for policymakers.
   
   
6. Scattered Presentation of Limitations:
   The limitations outlined in the Conclusions are robust, acknowledging constraints like the study’s scope and data limitations. However, similar limitations are mentioned in the Discussion, disrupting the flow of interpreting findings and reducing structural clarity. This scattering makes it harder for readers to assess the study’s boundaries systematically.
   
   
7. Literature Review Errors:
   The Literature Review mistakenly refers to the “Great British Fixed Link” instead of the “Great Belt Fixed Link,” introducing potential confusion about the study’s focus. Additionally, the abbreviation toggles between GBFL and BGFL without explanation.
   
   
8. Figure 4 Mislabeling:
   Figure 4, which presents predicted land use simulations, is mislabeled. The panels are labeled as (a), (b), (a), (b), despite the caption describing four distinct simulations (2010, 2024, 2030, 2036).

 

This study offers valuable insights into the socio-economic impacts of major infrastructure projects, and its findings have clear relevance for urban and regional planning. With careful revisions to strengthen the methodological transparency, structural clarity, and practical recommendations, the manuscript has strong potential to make a meaningful contribution to the literature.

Comments for author File: Comments.pdf

Author Response

We thank the reviewer for his/her constructive comments and suggestions, which have significantly helped us improve the manuscript. Below, we provide detailed responses to each question and point raised, indicating how we have revised the manuscript text or provided explanations for specific comments to the reviewers. All modifications are highlighted in the revised version of the manuscript.

____________________________________________

 

Comments and Suggestions for Authors

Review of "Assessing Mobility-Driven Socio-Economic Impacts on Quality of Life in Small Urban Areas: A Case Study of the Great Belt Fixed Link Corridor"

Thank you for allowing me to review your paper. It is an interesting and valid study, evaluatingthe impacts of the Great Belt Fixed Link (GBFL) on smaller urban areas through spatial modeling, surveys, and socio-economic analysis. The study’s findings are of interest because they highlight the GBFL’s role in revealing uneven socio-economic benefits. That being said, I have concerns with this study:

 

1. Isochrone Justification:
   The study’s scope is defined by a 45-minute commuting isochrone, but there is no justification for this choice, nor are sensitivity tests conducted with alternative isochrones (e.g., 30 or 60 minutes). This selection undermines the robustness as different isochrones could alter commuting patterns, urban growth estimates, and socio-economic impacts, limiting confidence in the results’ generalizability.

We did mention a 45-minute threshold, backed with references in a few places in the manuscript, but we now highlighted it very clearly in the 3. methodology section, under the 3.1 case study:

‘…based on travel time preferences identified in earlier studies conducted by the Danish Chamber of Commerce [25] and the Technical University of Denmark [26].’

However, to fully respond to the reviewers’ comment, I will mention that we use a 45-minute commuting isochrone based on the commuting preferences of Danes, supported by extensive mobility research conducted over various years in Denmark. According to a 2021 survey by the Danish Chamber of Commerce, mentioned in our manuscript, the majority of Danes are open to commuting up to 45 minutes. This aligns well with earlier findings from 2018, reinforcing the notion that this time frame represents a commonly accepted travel tolerance. Additionally, our project-specific survey again confirmed that respondents would be willing to accept commutes of 45–60 minutes when searching for new jobs, around the study region, GBFL. These insights are consistent with the well-known Zahavi’s Law and Marchetti’s Constant. In addition, we also performed a spatial analysis experiment for this project using ArcGIS and QGIS tools, confirming that a 45-minute isochrone by car covers a substantial portion of the study area (in this experiment, we did not consider rush hours, though). This establishes it as a practically relevant boundary for analysing accessibility and socio-economic impacts within the Danish context. However, we will, of course, also provide recommendations in the discussion for alternative travel distances of 30 or 60 minutes.

 

1. Missing ANN Model Parameters:
   The Artificial Neural Network model’s input parameters (e.g., number of hidden layers, learning rate, activation functions) are not provided, despite the model being central to the predictive analysis. These parameters are crucial for ensuring reproducibility.
   
   

We described those parameters and details for ANN in the methodology section now; however, to respond to the reviewers’ comments, we also mention it here that the ANN model used in this study was implemented through the QGIS software environment. The ANN architecture employed a standard feed-forward network consisting of one input layer, one output layer, and ten fully connected hidden layers. The model was trained using a random sampling method with a sample size of 8,000 pixels, distributed across the study area. To incorporate spatial context, a neighbourhood size of 1 was applied, meaning each sample considered a 3×3 pixel window. Training was conducted using a learning rate of 0.001, momentum of 0.001, ten hidden layers and a maximum of 200 iterations. These parameters were chosen to ensure a smooth learning curve while minimising the risk of overfitting. The activation functions used within the hidden layers were non-linear, enabling the network to capture complex spatial patterns. The model's performance was validated using Cohen’s Kappa coefficient and Pearson’s correlation coefficient.

 

 

1. Lack of Formal Hypotheses:
   The study poses broad research questions but lacks a formal hypothesis to guide the analysis. This makes the study appear descriptive rather than hypothesis-driven. It makes it harder to evaluate findings against specific expectations.
   
   

Thank you for this remark. The hypotheses are now integrated into the text.

In introduction: This study is grounded in the hypothesis that constructing the GBFL would enhance mobility and stimulate economic growth for cities on either side of the corridor. The assumption was that improved connectivity would promote labour market integration and spatial development.

In conclusion: These findings necessitate a revision of our initial hypothesis, which anticipated broader and more uniform economic gains from the GBFL. Although connectivity was improved as expected, the extent of socio-economic transformation was more limited and spatially concentrated than originally envisaged. 

 

1. Data Preprocessing and Temporal Misalignment:
   The data collection description lists sources but omits critical details on data preprocessing steps, and integration methods for geospatial and survey data. Additionally, using 2024 data for a 1990–2018 study introduces temporal misalignment without justification. One must infer the reason for the dates selected.
   
   

Thank you for this observation; it has now been described in section 3.2. and 3.3.

 

1. Superficial Analysis of Drivers and Policy Implications:
   The study identifies uneven socio-economic benefits across municipalities but provides a superficial analysis of the drivers behind these disparities (e.g., local economic conditions, toll costs) and fails to propose targeted policy interventions. Greater insights into the causes of outcomes and specific policies (e.g., subsidies for lagging regions) would enhance the study’s practical relevance for policymakers.
   
   

Our study aimed to investigate the long-term socio-spatial impacts of the GBFL on local municipalities, focusing specifically on accessibility, mobility patterns, and regional development trajectories. While we recognise the critical role that policy design, including toll schemes and fiscal mechanisms, plays in shaping infrastructure outcomes, our aim was not to prescribe specific interventions but to build a robust, evidence-based spatial analysis. Given the ongoing nature of policy debates and the diversity of stakeholder perspectives, we chose to frame our findings as a foundation for future dialogue rather than to offer definitive policy solutions. We hope that our findings contribute constructively to broader discussions on infrastructure planning, territorial equity, and inclusive mobility.

To address this limitation, we have added a small paragraph in section 5.4 indicating the need for future research.

 

1. Scattered Presentation of Limitations:
   The limitations outlined in the Conclusions are robust, acknowledging constraints like the study’s scope and data limitations. However, similar limitations are mentioned in the Discussion, disrupting the flow of interpreting findings and reducing structural clarity. This scattering makes it harder for readers to assess the study’s boundaries systematically.

Based on the review comments we no integrated few more comments in the discussion section to adress some other limitations as recommended, and we belive that now the text

1. Literature Review Errors:
   The Literature Review mistakenly refers to the “Great British Fixed Link” instead of the “Great Belt Fixed Link,” introducing potential confusion about the study’s focus. Additionally, the abbreviation toggles between GBFL and BGFL without explanation.

We have now identified the misspelling (Great British Fixed Link) and corrected it to Great Belt Fixed Link (GBFL).

1. Figure 4 Mislabeling:
   Figure 4, which presents predicted land use simulations, is mislabeled. The panels are labeled as (a), (b), (a), (b), despite the caption describing four distinct simulations (2010, 2024, 2030, 2036).

All the figures in the manuscript have now been changed; the resolution and size have been improved, and the enumeration has been removed as it is no longer necessary.

This study offers valuable insights into the socio-economic impacts of major infrastructure projects, and its findings have clear relevance for urban and regional planning. With careful revisions to strengthen the methodological transparency, structural clarity, and practical recommendations, the manuscript has strong potential to make a meaningful contribution to the literature.

 

Reviewer 5 Report

Comments and Suggestions for Authors

Irregularities in source citation and author identification occur which can be addressed via editing (e.g., lines 109 & 117).  Article numbers should not be used as subjects or objects in a sentence (e.g., lines 117 & 120 & 133 & 519). No attribution is given for the sources of each figure. Section 4.4 should be in methodology not in results. Results from 4.4 should then be in the results section. The mobility paradox definition is repeated unnecessarily in 5.2 (lines 498 & 509). Line 541 (and later 592), 2024 is indicated to be in the future rather than the past (it is 2025 now). If this exceeds the data timelines (trying to predict what 2024 should have been), this should be referenced as such rather than left as if 2024 was still a future possibility. These are all minor editing changes, however, and do not detract from the quality of the paper itself.

Author Response

We thank the reviewer for his/her constructive comments and suggestions, which have significantly helped us improve the manuscript. Below, we provide detailed responses to each question and point raised, indicating how we have revised the manuscript text or provided explanations for specific comments to the reviewers. All modifications are highlighted in the revised version of the manuscript.

________________________________

Comments and Suggestions for Authors

Irregularities in source citation and author identification occur which can be addressed via editing (e.g., lines 109 & 117). 

These irregularities have been resolved.

 

 

Article numbers should not be used as subjects or objects in a sentence (e.g., lines 117 & 120 & 133 & 519).

These irregularities have been addressed throughout the manuscript.

 

 

No attribution is given for the sources of each figure.

The figures are original, generated by the authors.

 

 

Section 4.4 should be in methodology not in results.

Section 4.4 has been revised, and a large part of the content has been moved to the methodology section.

 

 

Results from 4.4 should then be in the results section.

The results in this section still remain in 4.4, and it has been revised.

 

 

The mobility paradox definition is repeated unnecessarily in 5.2 (lines 498 & 509).

The repeated text has now been removed.

 

 

Line 541 (and later 592), 2024 is indicated to be in the future rather than the past (it is 2025 now). If this exceeds the data timelines (trying to predict what 2024 should have been), this should be referenced as such rather than left as if 2024 was still a future possibility. These are all minor editing changes, however, and do not detract from the quality of the paper itself.

Thank you for your valuable comment; we will correct the time related typo mistake.   

 

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for your efforts in revising the study. However, three issues remain unresolved. Below, I outline each issue with recommendations to enhance the paper’s rigor and contribution to the field.

1. Incomplete ANN Model Parameter Specification

Thank you for specifying ANN model parameters; however, the activation function remains only described as “non-linear,” which is insufficient for reproducibility, as specific functions (e.g., ReLU, sigmoid) significantly affect model performance (see Foody 2002). This outstanding issue hinders full replication, and I recommend explicitly stating the activation function used to ensure methodological transparency.

2. Repetitiveness Between Sections 5 and 6

Sections 5 and 6 are strikingly repetitive, with Section 6 echoing Section 5’s statistics on travel time reduction, public-sector job relocations to Odense and Ringsted, the ANN-CA model’s accuracy, and calls for richer migration data. This redundancy blunts the Conclusion’s impact, rendering it a second Discussion rather than a forward-looking synthesis, and it risks disengaging readers. I recommend reframing Section 6 as a concise, policy-oriented synthesis that (1) distills global lessons, (2) highlights actionable levers, and (3) proposes research agenda. Reserve methodological reflections concerns for Section 5 to restore narrative clarity.

3. Temporal Inconsistency

The study’s reliance on 2024 data (e.g., population, unemployment) in a 1990–2018 study period, without rigorous harmonization, undermines internal validity, as explanatory and response variables are not temporally aligned. While preprocessing for “temporal and spatial consistency” is mentioned, the paper lacks specific methods (e.g., scaling, interpolation) to align 2024 socio-economic data with 1990–2018 trends, risking biased coefficient estimates and misinterpretation of historical trends. To strengthen the study, I suggest three practical steps:

1. Justify the Use of 2024 Data: Explain why 2024 data was necessary and acknowledge limitations explicitly.
2. Clarify Harmonization Methods: Specify the exact method (e.g., “We scaled 2024 unemployment rates to 2018 using linear interpolation based on 2010–2018 trends, with weights derived from GDP growth”). Include a formula or reference a standard technique (e.g., Arellano-Bond adjustments).
3. Add a Robustness Check: Include a sentence or table showing results are consistent when 2024 data is excluded or adjusted differently. These adjustments will enhance the study’s credibility.

I look forward to seeing these revisions, which will strengthen your valuable contribution to urban mobility research

Author Response

First of all, we would like to sincerely thank the reviewer for taking the time to read our manuscript thoroughly, thereby helping to improve it. In the text below, we have tried to address all the comments provided in the second round of review. 

 

Comment: 

1. Incomplete ANN Model Parameter Specification

Thank you for specifying ANN model parameters; however, the activation function remains only described as “non-linear,” which is insufficient for reproducibility, as specific functions (e.g., ReLU, sigmoid) significantly affect model performance (see Foody 2002). This outstanding issue hinders full replication, and I recommend explicitly stating the activation function used to ensure methodological transparency.

 

Answer:  

We appreciate the reviewer’s observation and suggestion. Previously, we believed it was unnecessary and therefore omitted it, but we recognise the point and fully agree that specifying the activation function could be necessary for reproducibility and methodological clarity. Consequently, we have added a text in section 3.3 that reads:

‘Each hidden layer employed the sigmoid activation function, defined as σ(x) = 1 / (1 + e^−x), which introduces smooth non-linearity, bounds outputs between 0 and 1, and enhances gradient stability during backpropagation. These characteristics make it particularly suitable for modelling transition probabilities in spatial land use simulations [36] [37]. This activation function facilitates the introduction of smooth non-linearity and effectively outputs within the range of 0 to 1, rendering it particularly suitable for estimating transition probabilities within land use change modelling frameworks. Its differentiable form and bounded output range contribute to gradient stability during backpropagation, facilitating consistent convergence over multiple training iterations. Reflecting this…’

 

Comment: 

2. Repetitiveness Between Sections 5 and 6

Sections 5 and 6 are strikingly repetitive, with Section 6 echoing Section 5’s statistics on travel time reduction, public-sector job relocations to Odense and Ringsted, the ANN-CA model’s accuracy, and calls for richer migration data. This redundancy blunts the Conclusion’s impact, rendering it a second Discussion rather than a forward-looking synthesis, and it risks disengaging readers. I recommend reframing Section 6 as a concise, policy-oriented synthesis that (1) distills global lessons, (2) highlights actionable levers, and (3) proposes research agenda. Reserve methodological reflections concerns for Section 5 to restore narrative clarity.

Answer: 

We sincerely thank the reviewer for the very useful feedback. In response, Section 6 has been completely revised to avoid repetition, improve structure, and align with the suggestions provided by the reviewer. The reflections previously included in Section 6 are now discussed solely in Section 5 to preserve narrative clarity. We trust that this restructuring enhances the impact and readability of the manuscript.

 

Comment:  

3. Temporal Inconsistency

The study’s reliance on 2024 data (e.g., population, unemployment) in a 1990–2018 study period, without rigorous harmonization, undermines internal validity, as explanatory and response variables are not temporally aligned. While preprocessing for “temporal and spatial consistency” is mentioned, the paper lacks specific methods (e.g., scaling, interpolation) to align 2024 socio-economic data with 1990–2018 trends, risking biased coefficient estimates and misinterpretation of historical trends. To strengthen the study, I suggest three practical steps:

 

1. Justify the Use of 2024 Data: Explain why 2024 data was necessary and acknowledge limitations explicitly.
2. Clarify Harmonization Methods: Specify the exact method (e.g., “We scaled 2024 unemployment rates to 2018 using linear interpolation based on 2010–2018 trends, with weights derived from GDP growth”). Include a formula or reference a standard technique (e.g., Arellano-Bond adjustments). 
3. Add a Robustness Check: Include a sentence or table showing results are consistent when 2024 data is excluded or adjusted differently. These adjustments will enhance the study’s credibility.

I look forward to seeing these revisions, which will strengthen your valuable contribution to urban mobility research

 

 

Answer:  

In order to address all the points indicated below, we added a new paragraphs in section 3.3.

 

1. Justify the Use of 2024 Data: Explain why 2024 data was necessary and acknowledge limitations explicitly.

 

The inclusion of 2024 data was essential for simulating future land use patterns and assessing the prospective impacts of the GBFL on regional development. Since the study aims to inform planning scenarios extending to 2036, we incorporated near-future baseline data to serve as initial conditions for ANN-CA simulations. We now explicitly acknowledge it in Section 3.3

 

“Although these data fall outside the historical training window (1990–2018), their inclusion reflects a forward-looking, scenario-based modelling approach commonly used in spatial planning…”

We believe this justifies the use of 2024 data in our study.

 

 

1. Clarify Harmonization Methods: Specify the exact method (e.g., “We scaled 2024 unemployment rates to 2018 using linear interpolation based on 2010–2018 trends, with weights derived from GDP growth”). Include a formula or reference a standard technique (e.g., Arellano-Bond adjustments).

 

Thank you for this helpful comment. In response, we have revised Section 3.3 to explicitly describe the harmonization method used for the 2024 socio-economic variables (e.g., unemployment and population). Specifically, we used linear interpolation based on observed trends from 2010 to 2018 to ensure consistency with the historical training window. The text indicating these changes in the section is:

“… values were scaled using the formula:

 

X₍₂₀₂₄₎ = X₍₂₀₁₈₎ + ((X₍₂₀₁₈₎ − X₍₂₀₁₀₎) ÷ 8) × 6

Where X₂₀₂₄​ is the estimated value of a socio-economic variable for 2024, projected using a linear trend, X₂₀₁₀​ and X₂₀₁₈ ​are the observed values in 2010 and 2018, respectively. The denominator 8 represents the years used to compute the annual rate of change, and the multiplier 6 reflects the projection interval from 2018 to 2024. This trend-based adjustment aligns with standard forecasting techniques [38].  ….”

This method assumes constant annual growth, aligning with established pre-forecasting procedures commonly used in regional trend analysis. We have added a supporting reference to Hyndman & Athanasopoulos (2018) — Forecasting: Principles and Practice, which outlines these linear trend-based forecasting techniques. Again, this clarification has been included in Section 3.3 of the revised manuscript.

 

1. Add a Robustness Check: Include a sentence or table showing results are consistent when 2024 data is excluded or adjusted differently. These adjustments will enhance the study’s credibility.

Thank you for the suggestion. To ensure the reliability of our findings, we conducted an additional robustness test by re-running the ANN-CA model with the 2024 socio-economic predictors excluded. The results indicate only marginal variations in land use transition probabilities and Kappa scores (now mentioned in the text), confirming that the model’s predictive structure is not heavily dependent on the 2024 data. Instead of including the table, which does not significantly contribute to the new findings, we added a sentence in section 3.3, as the reviewer suggests. The sentence in section 3.3 is as follows:

“...To verify the model’s robustness, the ANN-CA simulation was recalibrated by excluding 2024 predictors; the results remained highly consistent (Kappa = 0.912, accuracy >97%), confirming that projections were not overly sensitive to assumptions about near-future data.”

This confirms that robustness was tested, satisfying the request for methodological rigour.