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

Abstract

Mobility is a fundamental catalyst for urban transformation, particularly in smaller urban centres, where enhanced transport can significantly influence socio-economic dynamics. This study investigates the socio-economic and spatial developments associated with the fixed-link transport corridor connecting the Zealand and Funen islands in Denmark. Despite its importance, a gap remains in understanding how this transport corridor has reshaped inter-regional connectivity over three decades and affected urban development in the surrounding areas. To address this gap, the study analyses the socio-economic effects of improved connectivity, focusing on residential relocation trends, commuting patterns, housing market dynamics, and employment in smaller communities. Adopting a mixed-methods approach that integrates surveys with spatial modelling, the analysis examines changes in commuting habits, economic opportunities, and land use from 1990 to 2018 within a 45 min travel radius of the corridor. The findings reveal that enhanced accessibility has widened commuting ranges, stimulated urban growth, and transformed housing and labour markets. However, these advantages are not evenly distributed, highlighting emerging spatial disparities. By merging empirical data with predictive models, this study enriches the discourse on sustainable urban mobility and spatial equity, providing valuable insights for policymakers and planners as they contemplate future fixed-link transport connectivity aimed at fostering inclusive regional development.

1. Introduction

Mobility is a fundamental pillar of contemporary urban systems, significantly influencing economic development, spatial integration, and overall quality of life [1,2]. Specifically, mobility-centric fixed-link connections reshape inter-regional commuting patterns, facilitate economic redistribution, and alter commuting patterns associated with inter-regional mobility [3,4,5]. The objective of fixed-link connectivity extends beyond merely reducing physical travel time, as it promotes the efficient and sustainable movement of individuals and goods across fragmented or insular geographies [2]. Nevertheless, despite the tendency to assess mobility-focused connectivity primarily through technical and economic frameworks, their long-term socio-economic and spatial impacts are frequently inadequately explored [4,5]. A prime example of such a strategic transport corridor is the Great Belt Fixed Link (GBFL), whose effects on society and the broader socio-economic landscape have not been systematically evaluated. This gap emphasises the necessity of robust and systematic methodologies to assess the impact of this significant transport corridor, ensuring informed decision-making and sustainable urban planning for similar connections in the future.

Denmark, despite its relatively modest geographical dimensions, consists of over 400 islands and is delineated by significant natural maritime divisions. Historically, these maritime barriers have presented substantial challenges to internal connectivity and national cohesiveness. The Great Belt Link serves as a pivotal example of a strategically developed transport corridor aimed at overcoming these obstacles. This corridor was envisioned to facilitate the integration of East and West Denmark into a cohesive mobility network, thereby reducing travel times and fostering both national and regional economic development. A comprehensive evaluation of mobility effects, undertaken by the Danish Ministry of Transport in conjunction with Sund & Bælt [6,7], primarily concentrated on traffic performance metrics and national economic benefits, particularly regarding diminished travel durations. However, the extensive socio-economic repercussions on the local population, including alterations in housing markets, variations in employment distribution, shifts in urban developmental patterns, and changes in commuting behaviours, remain insufficiently examined. 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.

Understanding these transformations profoundly is essential for evaluating the actual impact of such connectivity on regional equity and urban sustainability. In Denmark, fixed-link connections are closely linked with land use regulation through national planning instruments such as the Planning Act [8]. These tools are designed to manage urban expansion and mitigate the unintended consequences of mobility interventions, such as urban sprawl or spatial polarisation. Despite these mechanisms, critical questions persist: Does a fixed-link connection merely remove physical boundaries, or does it risk introducing new socio-economic ones? Can strategic connectivity facilitate inclusive development, or does it privilege certain regions and populations over others?

Consequently, this article addresses the above questions 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. Initially we hypothesised that the construction of the GBFL would enhance regional mobility and stimulate significant economic growth on both sides of the link. Utilising a robust array of methodologies, including spatial analysis, modelling, and survey data, the study elucidates the transformations in land use, commuting patterns, and housing market trends from 1990 to 2018. The primary objective is to critically evaluate the long-term impacts of mobility-oriented infrastructures on human migration and urban morphology while providing predictive insights pertinent to upcoming fixed-link initiatives, particularly the proposed Kalundborg–Aarhus corridor. By focusing on smaller urban areas significantly affected by the corridor, this research contributes to the academic discourse surrounding the socio-economic consequences of fixed-link connectivity, aiming to inform strategies for achieving equitable and sustainable urban futures.

2. Literature Review

2.1. Land Use and Mobility

Transport corridors play a pivotal role in shaping mobility patterns and spatial transformation. Numerous studies demonstrate that such infrastructure can redistribute populations, reshape land use dynamics, and redefine accessibility across urban and peri-urban regions [9,10]. Fixed-link corridors, such as bridges, introduce lasting shifts in human mobility and regional development trajectories [10]. Recent findings highlight that such structural assets not only improve travel efficiency but also foster new social and economic geographies [5,6]. Accessibility and affordability are key determinants of commuting choices, labour market participation, and residential preferences, all influenced by high-capacity transport infrastructure [11]. Evidence from the GBFL, as documented by the Danish Ministry of Transport and Sund & Bælt Holding, projects substantial gains in travel efficiency, inter-regional integration, and socio-economic performance [3,4,6,7].

The transformative effects of fixed links on land use are particularly evident in long-term Land Use and Land Cover (LULC) changes. Chu et al. [12], for instance, used 18 years of Landsat imagery and a machine learning classifier to assess the influence of fixed-link infrastructure on adjacent urban areas. By applying Interrupted Time Series Analysis (ITSA), the study isolated pre- and post-construction impacts, revealing strong correlations between the infrastructure and changes in tourism, freight activity, and regional economic growth. The research contributes to understanding the relationship between infrastructure development and land use dynamics, alongside its socio-economic implications. Similarly, a study by Cai et al. [13] on fixed-link corridors in disadvantaged mountainous regions investigated how such connections can simultaneously enhance accessibility and economic integration. Using spatial analysis and GIS scenario simulations, particularly the Jordà direct projection method, they estimated impulse response functions to explore the relationship between Gross Domestic Product (GDP) and public spending. Their results indicate that fixed-link connectivity spending positively influences economic output, although the effects vary by methodology. Notably, the study’s reliance on two socio-economic indicators highlights the necessity of more comprehensive data, including land use and industrial evolution, to fully capture these dynamics. The studies above indicate that fixed links such as the Great Belt have enhanced mobility, which essentially facilitates daily commuting patterns and contributes to integration and socio-economic gains [5,14]. However, these effects are not uniformly distributed across space. For example, Leduc et al. [15] argue that improved mobility does not automatically translate to balanced economic development, due to persistent spatial disparities, but requires coordinated spatial planning to ensure equitable outcomes. In order to support both retrospective and predictive analysis, spatial simulation tools such as Artificial Neural Networks, Cellular Automata, and hybrid GIS-based models have become increasingly important [12,16,17,18]. These methods offer planners a means of modelling future land use under different regulatory and demographic scenarios. Yet, as research highlights, the benefits of such infrastructure are often constrained by institutional factors like zoning regulations, planning inertia, and uneven regional capacities [5].

The studies discussed above highlight how fixed-link infrastructure alters land use and mobility patterns, frequently resulting uneven spatial accessibility. Nevertheless, the long-term spatial implications in smaller urban contexts, such as those influenced by the GBFL, remain to be investigated. This study expands upon these insights by employing predictive modelling to evaluate the impact of such infrastructure on commuting behaviour and urban form at a regional scale.

2.2. Socio-Economic Impact and Living Standards

Existing studies [4,19,20] indicate that living standards correlate with material factors such as income, GDP, life expectancy, and economic opportunity. While living standards and quality of life share similarities, quality of life emphasises aspects influencing well-being, including rights, freedom, and equality [21,22,23]. For instance, [21] argues that human development encompasses not only material conditions but also health, education, and cultural enrichment, while Costanza et al. [23] highlight how subjective well-being and material conditions together define spatial disparities in living standards. Despite this view, socio-economic research reveals various dimensions of living standards that enrich the socio-economic context of Denmark, as detailed in [24], illustrating connections to the economic structure and regional disparities. Similarly, Rallis et al. [14] emphasise that regional disparities in living standards are deeply influenced by spatial configurations, access to services, and planning capacity factors directly tied to infrastructure placement and accessibility.

Mulalic et al. [5], on the other hand, analysed the effects of the GBFL on productivity and labour market matching. They found that the improved accessibility led to notable gains in productivity and wages, particularly among firms employing highly educated males, while the wage impacts from labour market matching were less significant. This suggests that such infrastructure investments may disproportionately benefit specific socio-economic groups, exacerbating existing inequalities. Similarly, Yavuz et al. [25] quantified the economic impacts of the fixed-link corridor on local businesses, aiming to develop a systematic model that integrated established parameters from the literature to assess the economic implications of such connections for future stakeholders. Hybel et al. [26] investigated how transportation and urban amenities influence quality of life (QOL) in Danish housing markets, with a particular focus on commuting costs and household heterogeneity. Their model, based on the Rosen framework, found that access to transportation strongly correlates with housing prices and willingness to pay, especially among households with diverse needs. Using QOL indices derived from regressions of housing prices, wages, and commuting costs, [27,28] further confirmed that enhanced accessibility increases housing demand, but not uniformly. The findings underscore the critical role of transportation in shaping QOL and the importance of urban amenities in housing market evaluations. The research highlights that variations in household preferences and commuting costs significantly impact housing prices and QOL, while also addressing limitations and proposing further exploration of household heterogeneity and urban amenities. In another study, Leduc et al. [15] used the Jordà method to analyse the causal relationship between public spending and GDP. Their results confirm that infrastructure investments, including fixed links, can positively stimulate economic growth, though effects vary by methodology. Without inclusive planning, such infrastructure improvements risk reinforcing existing spatial and social inequalities. These insights underscore the importance of assessing not only the structural impacts of fixed links but also their distributive outcomes. This study addresses a gap through an integrated spatial and socio-demographic analysis of long-term regional transformations.

In summary, the literature demonstrates that while fixed transport corridors such as the GBFL can significantly enhance economic performance, productivity, and quality of life, these benefits tend to be unevenly distributed, often concentrated among specific socio-economic groups or regions. Without inclusive planning, such infrastructure improvements risk reinforcing existing spatial and social inequalities. Against this backdrop, urban researchers increasingly deploy spatial modelling tools to simulate land use dynamics and forecast socio-spatial change. This study contributes to that discourse by applying a sophisticated tool for the quantitative spatial analysis of LULC changes, integrating machine learning with GIS analysis. It supports historical land use, socio-economic trends, and the predictive modelling of LULC dynamics. The approach enables understandings of how fixed-link infrastructure reshapes regional development. Importantly, this tool is not a conclusive solution but serves as a framework to deepen our understanding of the behavioural and structural transformations caused by the GBFL. This study extends the existing literature on how fixed-link accessibility affects the urban form and spatial justice, uniquely combining predictive modelling with survey data on commuting attitudes to investigate both structural and experiential dimensions of mobility change.

3. Methodology

3.1. Case Study

The GBFL exemplifies a critical transportation connection that enhances inter-regional connectivity in Denmark (Figure 1). Beyond serving as a physical structure, it acts as a socio-spatial catalyst, altering mobility patterns and economic exchange between Funen and Zealand. To define the spatial impact of the GBFL, a 45 min commuting isochrone was developed through network-based spatial accessibility analysis using GIS, based on travel time preferences identified in earlier studies conducted by the Danish Chamber of Commerce [29] and the Technical University of Denmark, DTU [30]. This analysis identified municipalities within the daily mobility catchment area, facilitating an evaluation of socio-economic and land use implications. The study focused on ten municipalities within this threshold, which encompass a mix of urban and peri-urban environments, anticipated to undergo varying transformations due to improved accessibility and regional integration.

3.2. Data Collection

A robust combination of statistical, geo-spatial, and survey data was employed to conduct an in-depth spatial and socio-economic analysis of the GBFL’s impact on regional integration and mobility. These datasets were selected for their reliability, thematic relevance, and spatio-temporal resolution, facilitating both retrospective assessments and predictive modelling. The quantitative framework was based on datasets obtained from [31], the authoritative source for official statistics, providing standardised, high-quality data essential for comparative regional socio-economic assessments. Key indicators, including unemployment rates, population changes, and economic activity, were analysed to identify regional disparities in living standards and mobility potential. Additionally, supplementary socio-economic datasets were sourced from Nøgletal, an open governmental platform from the Ministry of the Interior and Health (ISM) which offers granular municipal-level indicators related to demographics, income, and employment, enhancing the understanding of local development conditions. Data from Boligstatistik, managed by Finance Denmark, provided crucial housing price indicators for analysing spatial dynamics in residential demand and investment. Integrating these diverse datasets (

Table 1. Data and data sources used for prediction model.

Name	Source	Year	Purpose
DAGI	Datafordeler	2023	Defining municipal boundaries for zonal statistics
Road network	Open Historical Map	2015–2024	Infrastructure development timelines and spatial interaction modelling
Commuting times	GIS OPS UG	2024	Isochrone modelling and accessibility analysis around the GBFL
CORINE	Copernicus	1986–2018	Land cover classification used as base for land use change modelling
Soil types	Aarhus University	1970	Environmental constraints in land use modelling
Creeks and streams	Geofabrik	2024	Hydrological features in spatial prediction models
Slope	GEBCO	2024	Physical geography input for land suitability modelling
Population	DST and ISM	2024	Demographic pressure and migration patterns
Unemployment rates	DST	2024	Socio-economic disparity and labour market accessibility
Housing prices	Finance Denmark	2024	Urban expansion and residential mobility trends

) enabled a comprehensive examination of the fixed link’s influence on economic flows, territorial cohesion, and lifestyle migration. These datasets were incorporated into a GIS-based analytical framework to support descriptive mapping and land use change prediction modelling, with CORINE land cover data playing a pivotal role in developing temporal layers for model training and scenario simulation. As the most recent validated CORINE dataset available at the time of the study was from 2018, spatial inputs for predictive modelling were harmonised to this temporal point to maintain consistency and ensure forecasting reliability (Table 1). All datasets underwent preprocessing to ensure temporal and spatial consistency. Geo-spatial data layers were standardised to a uniform coordinate system (ETRS89/UTM32N), and attribute tables were normalised for compatibility across various platforms. The data were filtered for completeness and restructured into thematic layers for integration within the GIS-based analytical framework. This preprocessing facilitated coherent input across both the modelling and descriptive components of the study.

To enhance quantitative analysis and capture subjective experiences related to the GBFL, an online survey was conducted among residents of municipalities within the 45 min isochrone travel zone. The survey assessed the perceived impacts of the fixed link on travel duration, employment accessibility, willingness to relocate for job opportunities, corridor utilisation frequency, and subjective regional quality of life. This primary data offers insights into behavioural responses and mobility choices, supporting the study’s aim of correlating network accessibility with socio-spatial transformation. By triangulating survey findings with spatial and socio-economic datasets, this research captures both the objective opportunities and subjective experiences of residents. The integration of these data sources enables comprehensive analysis, with statistical datasets providing essential context for regional socio-economic development and spatial layers aiding in land use forecasting. Moreover, the survey results contribute to a human-centred understanding of GBFL’s impact, ensuring alignment with research objectives that elucidate the spatial, social, and economic implications of fixed-link connectivity in Denmark.

3.3. Spatial and Predictive Modelling

To evaluate the spatial and socio-economic impacts of the GBFL on the regional development level, we developed a predictive land use change model using advanced Land Use Change Evaluation tools within the QGIS environment supported by Python 3.11 libraries. This model employed algorithms including Artificial Neural Networks (ANNs) [32], Logistic Regression (LR) [33], Multi-Criteria Evaluation (MCE) [34], and Weights of Evidence (WoE) [35]. LR, ANNs, and WoE utilise machine learning techniques to detect patterns in datasets, while MCE relies on expert assessments to discern key factors affecting land use change. In this study, the ANN was selected as the primary method due to its superior ability to model non-linear and high-dimensional relationships across spatio-temporal datasets. The ANN model employed a standard feedforward neural architecture with input, hidden, and output layers, and was trained using a learning rate of 0.001, momentum of 0.001, and ten hidden layers over 200 iterations. 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]. This activation function facilitates the introduction of smooth non-linearity and effectively produces 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, the model exhibited a smooth and stable learning curve, indicating effective convergence and minimised overfitting, thereby validating its capacity for generalisation and robust predictive performance.

The input data consisted of rasterised environmental and socio-economic layers, all normalised and spatially harmonised to align with the spatial extent and resolution of the CORINE Land Cover dataset (100-metre grid, EPSG:3857). Historical land use data from the CORINE programme (1990, 2000, 2006, 2012, and 2018) served as the ground truth for both training and prediction. Additional input layers included slope, soil type, and proximity to surface water, as well as the following infrastructural and socio-economic features: distance to major roads, historical and planned road networks, commuting zones, population figures, unemployment rates, and housing prices. To support consistency across temporal layers, all input rasters were clipped to a fixed spatial extent and harmonised across time steps, aligning with CORINE land cover time slices (1990–2018). To extend the predictive model beyond 2018, socio-economic variables for 2024 were incorporated to simulate land use scenarios through 2036. 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. To ensure internal consistency, all 2024 variables were harmonised through linear interpolation based on observed trends from 2010 to 2018. This procedure assumes linear growth and does not account for structural breaks or shocks post-2018, which are acknowledged as limitations of the extrapolation method. Specifically, values were scaled using the following 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, and 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 [37]. 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 the projections were not overly sensitive to assumptions about near-future data.

Isochrone modelling for commuting zones was conducted using the OSMnx 2.0.4 and NetworkX 3.5 Python libraries within QGIS 3.40.8 and Jupyter Notebook 7 environments. Commuting zones were generated using an OpenStreetMap-derived road network and constrained by a 45 min travel threshold [29], creating a binary raster mask that identifies labour market catchments. This network-based accessibility surface was utilised as an input feature in the land use model, facilitating the spatial representation of effective commuting corridors and GBFL-induced accessibility effects.

The training dataset was constructed from 8000 randomly sampled points across the study region, which included East Jutland, Funen, and the western and southern parts of Zealand. For each sample, a 3 × 3-pixel neighbourhood was considered, allowing the model to incorporate spatial context by analysing the focal pixel in relation to its immediate surroundings. To validate the model’s performance, land use was predicted for 2010 and compared with actual CORINE classifications. The results were assessed using Cohen’s Kappa coefficient [38,39], which quantifies agreement between two evaluators and provides a framework for assessing reliability. Pearson’s correlation coefficient [40] evaluated relationships between model inputs and land use transitions to ensure statistical independence among variables, thus supporting the model’s reliability in predicting scenarios from 2024 to 2036. Following the training and validation of the ANN [40], a Cellular Automata (CA) [16] simulation was utilised to extrapolate future land use dynamics for the years 2024, 2030, and 2036.

To reduce multicollinearity, variables with correlation values exceeding 0.5 were considered correlated and subject to removal. Specifically, population and unemployment exhibited seamless linear correlation (r = 1.0), while housing prices demonstrated moderate correlation with both population (r = 0.503) and unemployment (r = 0.503). To prevent overfitting and unstable weight updates during the ANN training process, unemployment and housing prices were excluded. This approach was grounded in the principle that redundant input features, those exhibiting high pairwise correlation, can introduce noise, inflate model complexity, and hinder the convergence process during training, even in non-linear models like ANNs. Although ANNs can capture non-linear patterns, they remain sensitive to the quality of the input space: high intercorrelation among variables can lead to overfitting, slower convergence, and reduced model interpretability. Removing such features promotes a more parsimonious input set, reducing dimensionality and enhancing generalisation. This process aligns with existing practices in spatial predictive modelling [17,40], where highly correlated predictors can result in convergent bias and compromised generalisation.

Following the training and validation of the ANN [40], a Cellular Automata (CA) [16] simulation was utilised to extrapolate future land use dynamics for the years 2024, 2030, and 2036. The CA model utilised transition probabilities derived from the ANN and employed rule-based iterations to simulate spatial diffusion. Operating on a 3 × 3 Moore neighbourhood, each pixel’s transition depended on the adjacent land use configuration and calculated probabilities. The CA were calibrated using observed land changes between 2012 and 2018, reflecting empirically derived transition dynamics.

The simulation outputs included classified land use maps and continuous certainty rasters, with the latter representing model confidence at the pixel level. For interpretability, outputs were consolidated into four land use classes: urban, industrial, infrastructure, and residual (undeveloped) land. Certainty values exceeded 99% across the simulation, indicating consistent predictive accuracy. Taken together, the integrated ANN-CA modelling framework provided a robust approach to simulating long-term land use transformations under enhanced regional connectivity conditions. This hybrid method effectively captured statistical and spatial drivers, offering valuable insights into development trajectories likely to follow the GBFL.

4. Results

4.1. Commuting Behaviour and Accessibility

According to the Danish National Survey [41], approximately 28% of daily transportation is commuting-related, predominantly using private automobiles for distances exceeding 5 km. While public transport exists, it plays a minor role for distances greater than 10 km. The average commuting distance in Denmark is around 22 km, with West and South Zealand commuters averaging 30.8 km, while those in Funen range from 25 to 29.9 km. In contrast, the average for Copenhagen’s metropolitan area is less than 15 km [31].

A shift in attitudes toward acceptable commuting times is evident, as data from the Danish Chamber of Commerce [18] reveals a decline in the willingness to commute up to 45 min from 43% in 2018 to 29% in 2021 [42].

This decline may reflect evolving preferences for work–life balance in the post-pandemic era. Although the observed trends support such interpretations, they remain hypothetical and necessitate further empirical validation through longitudinal or behavioural studies. The survey also highlights gender and educational disparities, noting that men and those with higher levels of education are generally more willing to commute longer distances, emphasising the need to consider socio-demographics when assessing infrastructure impacts on mobility.

The survey results suggest a strong correlation between fixed-link access and extended commuting behaviour, with over 60% of respondents willing to commute beyond 30 min due to improved accessibility. Many have relocated for affordable housing while maintaining employment in urban centres. The commuting time benchmark is critical for defining effective commuting catchment areas influenced by the GBFL. Network-based isochrone analysis has identified regions within a 45 min radius of the GBFL that benefit from enhanced labour market integration and residential flexibility. This improved access not only reduces travel times but also offers significant economic benefits, including increased labour mobility and potential reductions in CO₂ emissions through decreased reliance on ferry transport [43]. However, monitoring and managing potential challenges, such as increased traffic congestion and heightened accident risks, remain crucial in maintaining the reliability of the commuting corridor. This aligns with other studies showing that higher congestion levels correlate with increased crashes, indicating the need for effective safety management strategies [44].

4.2. Urban Growth and Land Use Change

The spatial analysis indicates a consistent pattern of urban expansion, especially influenced by the establishment of the GBFL, which has significantly impacted urban development patterns in municipalities within a 45 min commuting radius of the GBFL. This enhancement of infrastructure has improved inter-regional connectivity, thus facilitating increased mobility and accessibility between the islands of Funen and Zealand, likely due to a more robust integration with the Greater Copenhagen area. An examination of land use data spanning the period from 1990 to 2018 indicates a consistent expansion of urban areas within the affected municipalities. Urban land use experienced a notable increase of over 15% during this timeframe (Figure 2 and

Table 2. The table illustrates the number of pixels in the study area, along with the number of pixels representing urban areas within it. The proportion of urban area pixels is presented as a percentage to highlight differences in urban size across various time periods. This approach was similarly applied to the whole of Denmark, facilitating a comparison of percentages among the ten municipalities.

Year	Number of Pixels in Urban Areas	Number of Pixels in Municipalities	Urban Areas in Percentage	Number of Pixels in Urban Areas in DK	Number of Pixels in DK	Urban Areas in Percentage in DK
1990–2000	29,611	412,874	7.17%	289,143	4,292,914	6.74%
2000–2006	31,059	412,874	7.52%	303,530	4,292,914	7.07%
2006–2012	34,493	412,874	8.35%	324,130	4,292,914	7.55%
2012–2018	36,450	412,874	8.83%	343,190	4,292,914	7.99%

). From 1990 to 2018, the percentage of urban areas witnessed a rise both in the study area and nationally. Within the study area, urban land share expanded from 7.17% to 8.83%, representing a relative increase of approximately 23% over the 28-year period. Nationally, urban land share increased from 6.74% to 7.99%, resulting in a relative increase of approximately 18.5%. This suggests that municipalities bordering the GBFL experienced accelerated urban expansion compared to the national average. Notably, despite the adverse economic conditions stemming from the global financial crisis of 2008, urban growth continued at a moderate pace. This phenomenon is consistent with the findings of analogous studies, which posit that significant transportation, such as fixed-link connections, can catalyse urban expansion by enhancing accessibility and minimising travel times [16].

The observed urban expansion is closely intertwined with the state-led decentralisation of employment. As part of a broader spatial development strategy, several thousand government jobs were relocated from the Copenhagen metropolitan region to smaller municipalities in proximity to the GBFL. This relocation has undoubtedly played a pivotal role in urban growth. Approximately 37.6% of the 5750 individual state jobs relocated were situated within five of the ten municipalities in our study area, indicating a strategic decentralisation of employment opportunities and local labour market dynamics. This job redistribution is likely to have attracted population inflows, necessitating the expansion of urban areas to accommodate the influx of new residents. This relocation policy not only redistributed public-sector employment opportunities but also stimulated private-sector recruitment across a broader geographical catchment, particularly by enabling employers to access labour pools from both sides of the corridor. Enhanced commuting has consequently facilitated a broader functional labour market, potentially mitigating employment mismatches caused by geographic constraints.

Population data corroborates this trend, indicating substantial increases in municipalities like Nyborg and Ringsted following the opening of the GBFL. Nevertheless, we noted that population growth has not been uniformly distributed across all municipalities, implying that factors beyond enhanced accessibility, such as housing availability and local economic conditions, also exert significant influences on settlement patterns.

The development of the road network, particularly the completion of the highway configuration, has further enhanced regional connectivity. This comprehensive transportation framework has facilitated the more efficient movement of people and goods, thereby supporting urban growth and economic development in the connected regions.

In summary, the GBFL, in conjunction with strategic job relocations and road network enhancements, has played a pivotal role in shaping urban growth and land use changes in Denmark. These developments underscore the importance of integrated transportation and land use planning in promoting balanced regional development.

4.3. Socio-Economic Indicators

Evaluating the GBFL’s long-term socio-economic impacts necessitates looking beyond mobility metrics to include regional indicators such as employment and housing markets, which reflect changes in behaviour and structural transformations. While initial concerns centred on potential job losses in ferry-dependent towns like Nyborg and Korsør, these have been alleviated by evolving labour markets and property values, influenced by accessibility and economic factors. The analysis of unemployment data from 1992 to 2023 demonstrates that most of these municipalities experienced either a steady or accelerated decline in unemployment rates, frequently surpassing national trends. For instance, between 1992 and 2000, the municipalities of Kalundborg, Svendborg, Odense, and Næstved reported reductions in unemployment rates that exceeded the national benchmark of −4.3% [27]. Furthermore, during the economic downturn of the financial crisis from 2006 to 2012, the municipalities of Ringsted, Slagelse, and Kalundborg exhibited a more resilient response to the recession than the national average, with significant rebounds occurring in the post-crisis period from 2012 to 2018. These trends align with the relocation of jobs as a decentralisation strategy, and a substantial number of these relocated positions were established in municipalities within the 45 min commuting isochrone, specifically Odense, Slagelse, Sorø, Næstved, and Ringsted, which may account for the localised employment gains found in the research. However, the observed decline in population growth in these regions indicates that the demand for labour was predominantly met by the existing residential population, facilitated by improvements in commuting infrastructure rather than through migration. This challenges assumptions in spatial equilibrium theory, where job relocation is typically expected to attract population movement. Instead, the GBFL appears to have expanded the spatial reach of labour markets, enabling individuals to maintain their residence while accessing a broader range of job opportunities, an effect consistent with recent studies on infrastructure-led commuting corridors [45].

Parallel to labour market shifts, the housing market provides further insight into the socio-spatial effects of the link. An analysis of housing prices across ten municipalities accessible by the corridor from 1992 to 2024 indicates an overall upward trend [27,28], with variations across municipalities and time periods (Figure 3). Between 1992 and 2000, shortly after the GBFL’s approval, housing prices in Denmark increased by 86.36%, with municipalities like Nyborg and Næstved exceeding the national average. This initial growth suggests that expectations of future benefits were likely already incorporated into local housing markets, aligning with findings on the anticipatory effects [46]. Subsequent years showed similar variability; for instance, between 2000 and 2006, Odense, Slagelse, Næstved, and Kalundborg surpassed national housing growth rates. Conversely, during the financial crisis from 2006 to 2012, national housing prices declined nationally (−18.87%) and in all municipalities [28], with a slower recovery in West Zealand. By 2024, while the national average had reached pre-crisis levels (16,777 DKK/m²), municipalities in West Zealand such as Sorø, Slagelse, and Ringsted had not fully regained 2006 price levels.

In contrast, Odense, Svendborg, and Kalundborg demonstrated significant post-crisis recoveries, likely driven by urban infrastructure investments, university expansions, and job decentralisation. The Funen region generally outpaced West Zealand’s recovery, where persistent housing affordability persists at a higher level but growth potential is diminished. We observed a correlation between population trends and housing demand; while most municipalities experienced limited population growth, Ringsted exhibited an exception, suggesting a link between demographic expansion and housing demand. While the corridor improved access, it did not uniformly boost residential influx or price growth, highlighting the need for infrastructure to align with demographic dynamics, economic diversity, and planning strategies for sustainable market effects.

4.4. Land Use Prediction Modelling and Future Urban Expansion

According to the results, the model yielded a Kappa coefficient of 0.9248, placing its predictive performance in the “almost perfect” agreement category, and 98% spatial accuracy when validated against the CORINE 2012 land use dataset. The land use prediction for 2010 indicated modest yet noticeable urban expansion, with urban land (cities, industries, and infrastructure) increased by 0.34%, matched by a proportional decline in undeveloped land (Figure 4). The transition matrix indicated strong land use persistence, with over 99% of raster cells remaining in the same category between 1990 and 2000, a pattern consistent with gradual urbanisation and long-term land conversion dynamics.

The validated model was subsequently used to project future land use for the years 2024, 2030, and 2036, employing CORINE datasets from 2012 and 2018, as the most recent comprehensive data available. This strategy was adopted to ensure input consistency, and all spatial drivers (slope, population, etc.) were harmonised to align with the 2018 data layer, thereby ensuring uniformity through the analysis. The input variables were again filtered for multicollinearity; e.g., population and housing prices showed a correlation of r = 0.935, prompting the removal of the latter. The updated model returned a Kappa of 0.961, demonstrating highly reliable predictive performance. The simulation results indicated continued but slow-paced urban expansion, concentrated in areas already experiencing growth. Urban categories (cities, industries, infrastructure) gained marginal area over each period (e.g., ~0.09% growth in cities between 2012 and 2018), while undeveloped land declined incrementally (Figure 4). Certainty maps revealed the model confidence to be >99% across the study area, suggesting that spatial change is predictable and tightly linked to underlying physical and socio-economic drivers. It is remarkable that the comparison between CORINE 2000 and the simulated 2024 output revealed meaningful cumulative change: urban areas (cities) increased by approximately 481.56 km² (0.97%), while industrial zones expanded by 137.60 km². These values obviously demonstrate that although change is subtle at the pixel level, it reflects a long-term urban growth trajectory.

Notably, the simulations for 2030 and 2036 exhibited no detectable significant expansion beyond 2024. This suggests either model stabilisation, reflecting a saturation point in spatial growth patterns, or limitations in the data resolution and training span (2012–2018) in capturing small-scale but realistic urban development increments over short intervals. Moreover, the CA algorithm used in the model simulates land use change only when transition potential thresholds are exceeded. When calibrated on historical data with low transition intensity, such as the 2012–2018 CORINE dataset, the CA framework inherently preserves spatial continuity, producing minimal change across subsequent forecast iterations. These findings align with established patterns of urban inertia, wherein land use transformation progresses incrementally within highly regulated and spatially constrained regions like Denmark.

5. Discussion

5.1. GBFL as a Mobility Enabler

The GBFL has significantly transformed Denmark’s spatial dynamics by reducing both physical and perceived distances between Zealand and Funen. This substantial infrastructural investment has facilitated a reconfiguration of daily mobility thresholds, as demonstrated by the expansion of viable commuting zones, particularly within the 45 min isochrone. This observation aligns with established theories of time–space compression [47,48], which suggest that advancements in transportation technology reduce the temporal barriers associated with spatial traversal, thereby reshaping urban forms and labour geographies. From a practical perspective, the GBFL notably extended the functional urban region, facilitating new opportunities for interactions within the labour and housing markets. Although travel times have been significantly reduced from 60 min via ferry to just 12 min by car, empirical survey data and commuter statistics indicate that the propensity to commute remains inherently limited, often impeded by behavioural preferences and economic considerations. According to the Danish Chamber of Commerce [18], only 29% of participants expressed a willingness to undertake commutes exceeding 45 min each way, thereby highlighting the persistent friction of temporal constraints despite advancements in transportation efficiency. While the GBFL supports new commuting patterns, structural and economic barriers, such as the toll, continue to serve as selective filters, determining who benefits most from improved connectivity. This duality of access versus exclusion is a crucial aspect of mobility frameworks [49,50] and must be considered in the evaluation of large-scale infrastructure as more than mere technical artefacts. Furthermore, the current study does not incorporate in-depth qualitative insights into individual behavioural adaptations to new infrastructure. Future research should investigate how mobility choices and lifestyle adjustments are made in response to evolving transport systems, possibly through longitudinal or ethnographic methodologies.

5.2. Socio-Spatial Disparities and Urban Planning

Despite improved connectivity, the benefits of the corridor are unevenly distributed. Housing affordability has declined in some regions, and employment gains are spatially clustered. This reflects the “mobility paradox”: improved accessibility does not automatically result in equitable development unless supported by proactive planning and institutional coordination [51], which is consistent with findings in [52,53]. The primary research question addressed whether enhanced regional connectivity contributes to a more balanced territorial development pattern. The findings indicate that, although infrastructure expansion has created new opportunities, the resulting socio-economic benefits manifest in a spatially concentrated and uneven manner. For example, municipalities such as Odense, Ringsted, and Slagelse displayed relatively stronger demographic and employment indicators throughout the study period, including reduced unemployment rates and moderate population growth. However, these developments were not consistently observed across the region, indicating spatial concentration rather than uniform regional effects [27,28].

The transfer of 5750 state-sector positions to peripheral municipalities appears to have led to a reduction in unemployment; however, this change has not been met with a significant increase in residential migration [3,27,54]. This observation indicates that job relocation, in isolation, is inadequate for fostering meaningful spatial rebalancing. This is partly attributable to structural factors such as housing market rigidity, institutional asymmetries, and the limited capacity for smaller municipalities to absorb growth. Constrained housing supply, variations in local planning capacity, and unequal distributions of public investments can hinder the equitable realisation of accessibility benefits. These structural limitations underscore that infrastructure alone cannot deliver spatial equity unless accompanied by comprehensive socio-institutional reforms.

These findings are generally consistent with previous studies [52,55,56,57,58], 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.

Interestingly, the post-fixed-link period did not generate the level of urban transformation anticipated in earlier planning discourses. One explanation, supported by research in [59], suggests that individuals are more inclined to change jobs rather than relocate when confronted with increased commute times, emphasising the behavioural rigidity of residential choice. This challenges classical assumptions in urban settlement theory and highlights the necessity for behaviourally realistic models in regional planning. Additionally, responses from surveys and media discourses indicate that the corridor is still widely regarded as a financial obstacle, with toll charges disproportionately impacting lower-income demographics [54]. This prevailing perception poses a risk of solidifying territorial disparities, thereby undermining the infrastructure’s intended function as a facilitator of unity. However, since this research relied primarily on net population changes, we recommend a more granular migration flow analysis to trace actual mobility patterns. Although, considering imminent initiatives such as the Kattegatforbindelsen, these findings underscore the importance of incorporating equity-sensitive transport pricing and comprehensive accessibility planning.

5.3. Predictive Planning and Future Fixed Links

The other objective of this study was to assess the applicability of predictive modelling tools in simulating future urban expansion and supporting planning policies. Using an ANN combined with CA, the model demonstrated a high level of predictive capability for urban land use change, achieving a spatial accuracy of 98% and a Kappa score of 0.92, leading to the robust performance of the method. These validation metrics confirm the model’s reliability in replicating historical land use patterns, supporting earlier findings that ANN-CA frameworks are well-suited for land use forecasting in spatial planning contexts [12,16,17,40]. However, the forward simulations for 2024, 2030, and 2036 indicate no substantial urban expansion beyond 2024. This unanticipated plateau likely reflects the gradual and path-dependent characteristics of urban transformation in Denmark, where land use transitions occur incrementally due to regulatory constraints and demographic stability. Alternatively, it may suggest that the model’s sensitivity to low-transition training data from 2012 to 2018 may have contributed to a saturation effect, pointing to the need for more temporally detailed and transition-rich datasets in future modelling efforts.

The predictive tool holds significant promise as a planning support. According to Danish municipal legislation, municipalities are required to formulate 12-year expansion strategies, which must be substantiated with evidence-based projections. The model offers a robust, scientifically grounded framework for simulating spatial development under varying conditions. It integrates inputs such as accessibility, demographic trends, soil characteristics, hydrological factors and more (Table 1), enabling municipalities to explore multiple land use scenarios. These capabilities ensure compliance with national planning guidelines on preventing urban sprawl and provide a defensible basis for zoning and land conversion decisions. Moreover, the model’s spatial weighting flexibility allows planners to prioritise locations based on policy objectives, whether for densification, green infrastructure, or transit-oriented development. These features correspond with the current policy emphasis on data-driven planning, sustainability, and strategic foresight in urban governance. Nevertheless, further integration of such predictive tools into practical planning processes is necessary, particularly regarding the transparency of assumptions and the iterative involvement of stakeholders.

The GBFL highlights the broader role of mobility-oriented infrastructure in shaping regional development. Yet, the influence of such projects is mediated by contextual factors, including regulatory frameworks, institutional capacity, and the behavioural contexts in which they operate. Predictive models and data-intensive tools enhance planning capabilities but must be contextualised within the concrete realities of social disparities, regional capacities, and political economies. As countries advance their fixed-link connections through initiatives such as the Kattegat corridor, the insights gained from the Great Belt experience provide an essential understanding: infrastructure does offer a range of opportunities, the realisations of which are intrinsically linked to inclusive design, equitable governance, and forward-thinking planning.

5.4. Methodological Considerations and Future Research Directions

Although the study presents a robust spatio-temporal analysis of the socio-spatial impacts of the GBFL through advanced modelling, several methodological limitations should be acknowledged to contextualise the results and guide future research. Most notably, the reliance on raster-based land cover data (CORINE, 1990–2018) constrains the resolution at which urban transformations can be detected, particularly in small or peri-urban settlements where changes often occur at sub-pixel scales. Furthermore, the absence of high-resolution post-2018 land use and commuting data limited the model’s ability to capture more recent dynamics in behavioural shifts. While the ANN-CA framework effectively models land use transitions and captures structural drivers, it does not account for qualitative dimensions such as perceived accessibility, social equity, or lived experience. These aspects are essential for fully understanding the human response to infrastructure investments. Future studies would benefit from incorporating qualitative data sources, such as interviews, ethnographic fieldwork, or longitudinal panel surveys, to complement the quantitative spatial model and better interpret community-level adaptation to change.

An important limitation of this study is its reliance on net population change as a proxy for demographic dynamics. This aggregate metric obscures intra-regional and inter-municipal migration flows, which are crucial for understanding patterns of residential mobility and socio-economic restructuring. While register-based migration data were not accessible for this study, their inclusion in future research would enable more granular, spatially sensitive insights into population redistribution processes, particularly those influenced by toll costs, accessibility improvements, and local housing markets. In particular, demographic redistribution under infrastructure transformation is rarely uniform. Consequently, the current reliance on municipal-level aggregates may mask the local impacts of infrastructure investments. More refined migration flow datasets, e.g., origin–destination pairs or mobility data, are essential for capturing the directionality and intensity of migration patterns, which are vital for accurately forecasting land use and housing pressures. This concern becomes especially pertinent when interpreting the varied socio-economic responses observed across municipalities connected by the GBFL.

Moreover, while this study includes attitudinal survey data to assess perceived mobility and accessibility changes, it lacks deeper qualitative methods that could provide richer insights into behavioural drivers. For instance, comprehending how toll burdens, reductions in travel time, or perceived access to opportunity influence household decisions to relocate or remain would necessitate more extensive data collection. A mixed-methods design, incorporating ethnographic accounts and longitudinal interviews, would enhance interpretation and policy relevance. These limitations also extend to the analysis of spatial inequality. Although our findings reveal persistent disparities in development outcomes despite improved connectivity, the study does not adequately address the structural mechanisms behind these disparities. Constraints such as land use regulations, limited housing supply, or uneven institutional support may impede the equitable distribution of the benefits of transport infrastructure. Tackling these factors would require dedicated analytical frameworks drawing on official policy analysis, housing data, and governance structures to uncover the complex interplay between planning, regulation, and socio-spatial outcomes.

Given these gaps, we refrain from issuing detailed policy suggestions within this study. However, we emphasise that future research could explore targeted policy tools, such as mobility subsidies for low-income commuters or the spatial redistribution of public services. These measures, which are increasingly debated in infrastructure planning, may help mitigate the uneven impacts observed across the GBFL region. Additionally, the aggregated nature of several datasets, particularly at the municipal level, limits the detection of finer-grained variation in land use and behavioural responses. Household-level data, for example, would be better suited for analysing relocation decisions and transport mode adaptation. Without such granularity, there is a risk of overgeneralising findings and underestimating spatial heterogeneity. To overcome these constraints, future research should embrace multimodal data integration. Mobile phone tracking data, anonymised GPS logs, and smart card data can significantly enhance the spatio-temporal resolution of mobility analysis. When combined with detailed household or commuter panel surveys, these methods can illuminate short-term relocation patterns, motivations, and constraints. Furthermore, cross-regional comparative studies would provide valuable context, enabling researchers to benchmark infrastructure effects across diverse socio-political, economic, and planning environments.

In summary, while this study contributes to understanding the long-term socio-spatial effects of large-scale fixed-link corridors, advancing this field necessitates richer, more granular data and the integration of qualitative perspectives. Such enhancements would facilitate a more holistic exploration of how infrastructure interacts with governance and social behaviour to shape territorial development and equity.

6. Conclusions

As discussed in Section 5, this study has examined the long-term socio-spatial impacts of the GBFL through a comprehensive methodology that integrates spatial analysis, socio-economic datasets, and predictive land use modelling, utilising ANN–CA. The findings indicate that while the GBFL has substantially decreased travel times and enhanced inter-regional accessibility [3,54], its developmental effects exhibit notable spatial disparities and are significantly influenced by broader structural and institutional factors. A significant global insight derived from the GBFL case is that infrastructure alone cannot ensure equitable territorial development [5,23]. Connectivity acts as an enabler rather than a determinant, and the capacities of local governance shape its benefits, the robustness of planning frameworks, and the behavioural responses of the community. In the absence of complementary interventions, extensive transport infrastructure is likely to reinforce existing socio-economic patterns instead of promoting their redistribution [27,28].

To effectively translate accrued accessibility gains into equitable development outcomes, infrastructure planning should be explicitly embedded within comprehensive socio-spatial strategies. First, the implementation of context-sensitive pricing mechanisms, such as toll subsidies aimed at low-income commuters, has the potential to alleviate existing access barriers [60]. Second, municipalities should adopt responsive zoning and housing strategies that align with evolving accessibility conditions, thereby ensuring the provision of affordable housing near emerging mobility corridors. Third, it is essential to institutionalise scenario-based modelling tools, such as ANN–CA, within planning workflows to simulate alternative development trajectories, thus facilitating more transparent and evidence-based decision-making processes. Despite the study’s robust methodological framework, it is ultimately constrained by several significant limitations. In particular, the reliance on aggregated population and commuting data obscures intra-regional variations and behavioural nuances. Furthermore, the commuting threshold employed within the spatial scope of this analysis may inadequately encompass longer-distance functional linkages or migration responses. Addressing these gaps requires the integration of high-resolution mobility data, including origin–destination matrices, GPS data, and mobile phone records, to better reveal commuter behaviours and the dynamics of residential mobility.

Future research should emphasise the inclusion of qualitative methodologies, such as ethnographic fieldwork and structured interviews, to achieve a comprehensive understanding of how infrastructure affects everyday life, housing choices, and perceptions of opportunity. Additionally, conducting comparative analyses across various infrastructural contexts, including the forthcoming Kattegat Link, would significantly enhance the generalisability of findings by clarifying how different governance systems and spatial planning traditions influence the effects of fixed-link infrastructure [8]. In conclusion, the GBFL exemplifies the capacity of mobility infrastructure to improve regional accessibility. However, the equitable realisation of this potential depends on strong institutional coordination, inclusive spatial planning, and strategic data-informed policy interventions. The success of infrastructure-led development ultimately depends not only on enhanced physical connectivity, but also on cross-sectoral governance, integrated spatial planning, and socially inclusive design. Therefore, the experience of the GBFL provides critical policy-relevant insights into how infrastructure can operate not merely as a technical intervention, but as a strategic instrument for fostering territorial cohesion, advancing spatial justice, and promoting sustainable regional development.