Abstract
In a global context of raising urbanization rates, the study of land use and land cover changes (LULCC) has gained growing importance. Catalonia necessitated a comprehensive analysis of land use alterations for sustainable planning due to elevated urbanisation rates and infrastructural demands. This research investigated the relationship between LULCC and transportation infrastructure networks (TINs) in Catalonia from 2006 to 2018 at municipal level. Land use transformation, and transportation planning are deeply interconnected, requiring strategies that balance the urban needs with environmental conservation. In this setting, understanding the complex connections between land use changes and transport infrastructure proved essential for developing solutions that tackle urban growth and sustainability. Using a robust dataset that includes land uses from Corine Land Cover (CLC) and transport networks from OpenStreetMap (OSM), the research assessed how transportation infrastructure is correlated with urban expansion, employing advanced spatial analysis and a generalized least squares (GLS) regression model. The study highlights the need for smart land-use allocation to promote sustainable growth, revealing the impact of factors such as capital city status and proximity to transportation infrastructure, particularly seaports and airports. Results indicates that urban saturation in capital cities limits additional growth, whereas areas next to seaports and airports exhibit considerable expansion in artificial surfaces. These findings underscore the significance of cohesive land-use planning that harmonizes urban growth with sustainable development objectives.
Keywords
Land use change, transport infrastructure, Catalonia, Corine Land Cover
Introduction
The rising rates of global urbanization have made the examination of land use and land cover changes (LULCC) a vital field of inquiry, as land use patterns can have diverse economic, social, and environmental impacts (Litman, 2023). In dynamic regions like Catalonia, where urban expansion occurs rapidly, understanding the impact of transportation infrastructure networks (TINs) is essential for sustainable planning. On the one hand, the development of TINs can improve accessibility and in turn increase the demand for more urban development. On the other hand, the urbanisation of land can result in growing demand for transport and an increase in the demand for TINs (Dena Kasraian,2017). This study directly investigates the correlation between TINs and land cover changes from 2006 to 2018 at the municipal level in Catalonia.
Despite prior studies highlighting the impact of factors such as population growth and political decisions on land use patterns, is a knowledge gap in thoroughly understanding the specific correlation between transport infrastructure and land cover changes during the last decades in Catalonia.The research uses a comprehensive dataset that includes land use data from the Corine Land Cover (CLC) database to examine the relationship between transport infrastructure and the expansion of total artificial surfaces (TAS)1. This study utilises advanced spatial analysis techniques with GIS and a generalised least squares (GLS) regression model. The results show how transport infrastructure, particularly proximity to seaports and airports, is linked to urban expansion and emphasizes the role of capital city status and accessibility in shaping growth patterns.
Data gathering and variables’ preparation
The primary data were sourced from the Corine Land Cover (CLC) database, concentrating on land cover alterations from 2006 to 2018 in Catalonia. Information regarding transport networks, encompassing highways and train stations, was obtained from OpenStreetMap (OSM). Given that Catalonia has only four principal airports, we manually incorporated their locations into GIS2 utilising the digitisation function. Likewise, port data from the EEA SDI3 platform were utilised, with distances to ports and other infrastructures calculated using the 'Near'4 operation in GIS tools.
To capture geographical and political influences, we identified administrative capitals and those municipalities located along the coastline corridor. We transformed municipalities into points utilising the 'Feature to Point5' tool and computed the distance from each municipality to the nearest infrastructure (e.g., ports and airports) employing the 'Near' command with the algorithm configured to 'Geodesic6'. We imported highway and municipality data into GIS, conducted a spatial join 7to link municipalities with nearby highways, and analysed accessibility. The ETRS 1989 LAEA8 projection system was used to ensure spatial accuracy in analysing land cover and infrastructure data across Catalonia.
Map 1 illustrates the spatial distribution of built-up area changes (TAS2006 vs. TAS2018) and their relationship with transport infrastructure across the study area. We utilised the 'Select by Location' 9feature to discover municipalities adjacent to the sea and subsequently performed a spatial join to incorporate shoreline data with the selected municipalities.
The analysis evaluates two factors: the accessibility of infrastructure in each municipality, modelled as a dummy variable indicating presence or absence of access, and the proximity of each municipality to the nearest infrastructure. A supplementary variable encompasses available land to evaluate accessibility and expansion potential. These methodologies facilitated a thorough assessment of transport infrastructure's relationship with land use and land cover change, improving the spatial understanding of development trends.
Figure 1. spatial distribution of built-up area changes (TAS2006 vs. TAS2018) and their relationship with transport infrastructure across the study area.
Regression model
This study used a generalized least squares (GLS) regression model. A GLS regression is an enhancement of the standard linear regression model that corrects for heteroscedasticity, hence providing more efficient and unbiased coefficient estimations. The model employs robust standard errors to address heteroscedasticity, which may result from the diversity in factors affecting land type growth among municipalities. This method is selected due to the existence of varied geographical and infrastructural attributes that may result in non-constant variance in the error terms. Consequently, a standard ordinary least squares (OLS) regression may yield inefficient and possibly deceptive conclusions. The robust method enhances the reliability of statistical conclusions by addressing these heterogeneities, hence improving the model's applicability in situations where typical assumptions regarding error variance may be invalid.
Empirical specification
The empirical model employed in this analysis aims to explain the growth of Total Artificial Surface (TAS) between 2006 and 2018 in Catalan municipalities, using a set of explanatory variables that account for infrastructural and geographical factors.
The dependent variable, ln_TASgr, denotes the natural logarithm of the change in TAS over the study period. The model incorporates lnTAS2006, the natural logarithm of TAS in 2006, to account for the initial degree of urban fabric and to correct for baseline effects. Principal explanatory variables comprise various dummy variables: Station (1 if the municipality possesses a railway station, 0 otherwise), Capital (1 if the municipality serves as the regional capital, 0 otherwise), PortS (1 if the municipality contains a seaport, 0 otherwise), Airport (1 if the municipality has an airport, 0 otherwise), and Coastline (1 if the municipality is situated on the coast, 0 otherwise). Supplementary variables include the natural logarithms of distances to essential infrastructure: lnRailwayStation, lnHighway, lnAirport, and lnPort, indicating accessibility to transportation and trade networks. The model incorporates lnfreesoil06, the natural logarithm of accessible land in 2006, to address potential expansion limitations. A constant term is incorporated to account for unobserved variables influencing TAS growth.
Results
This section delineates the findings of our investigation, encapsulated in Table 1, which displays six distinct models estimated by generalised least squares (GLS) regression with robust standard errors. The models were designed to sequentially integrate distance factors to mitigate potential multicollinearity and more effectively isolate the impact of each distance measure. Model 1 functions as a baseline, devoid of any distance factors. Model 2 incorporates the distance to the nearest railway station, Model 3 contains the distance to the nearest highway, Model 4 accounts for the distance to the nearest airport, and Model 5 integrates the distance to the nearest seaport. Ultimately, Model 6 incorporates all distance variables concurrently.
Table 1. Generalized Least Squares (GLS) Regression Results on Total Artificial Surface Growth (2006–2018)
| (1) | (2) | (3) | (4) | (5) | (6) | |
| VARIABLES | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 |
| lnTAS2006 | -0.104*** | -0.111*** | -0.111*** | -0.110*** | -0.108*** | -0.118*** |
| (0.018) | (0.019) | (0.019) | (0.019) | (0.018) | (0.020) | |
| Station | 0.268 | -0.605 | 0.181 | 0.203 | 0.065 | -0.462 |
| (0.334) | (0.447) | (0.353) | (0.345) | (0.364) | (0.457) | |
| Capital | -2.168*** | -4.778*** | -3.993*** | -2.748*** | -1.958*** | -5.273*** |
| (0.401) | (1.108) | (1.232) | (0.307) | (0.610) | (1.541) | |
| PortS | 0.619* | 0.860* | 0.834*** | 0.527** | -0.956 | 0.017 |
| (0.355) | (0.449) | (0.277) | (0.216) | (0.845) | (0.825) | |
| Airport | 0.929** | 0.774 | 1.061** | -0.886 | 0.571 | -0.185 |
| (0.443) | (0.518) | (0.519) | (0.814) | (0.657) | (0.924) | |
| Coastline | 0.159 | 0.102 | -0.031 | 0.119 | 0.033 | -0.105 |
| (0.107) | (0.140) | (0.164) | (0.119) | (0.124) | (0.189) | |
| lnfreesoil06 | -0.123 | 0.046 | -0.065 | -0.043 | -0.080 | 0.085 |
| (0.142) | (0.155) | (0.139) | (0.137) | (0.143) | (0.148) | |
| lnRailwayStation | -0.473*** | -0.282 | ||||
| (0.181) | (0.227) | |||||
| lnHighway | -0.281* | -0.210 | ||||
| (0.136) | (0.206) | |||||
| lnAirport | -0.675** | -0.344* | ||||
| (0.274) | (0.185) | |||||
| lnPort | -0.692** | -0.378 | ||||
| (0.279) | (0.321) | |||||
| Constant | 0.419 | 0.711* | 0.464 | 2.579** | 3.099*** | 3.190** |
| (0.343) | (0.367) | (0.352) | (1.043) | (1.142) | (1.513) | |
| Observations | 947 | 947 | 947 | 947 | 947 | 947 |
| R-squared | 0.0703 | 0.0772 | 0.0733 | 0.0747 | 0.0753 | 0.0817 |
| Source: Authors. Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 | ||||||
The results indicate that the natural logarithm of Total Artificial Surface in 2006 consistently has a negative and significant impact on artificial surface expansion in all models. This pattern corresponds with urban densification theories, indicating that regions with higher starting urbanization levels exhibit reduced expansion rates, signifying urban saturation. Municipalities exhibiting elevated initial levels of urban development show diminished growth rates, underscoring geographical limitations and less demand for new construction in densely populated regions. This negative connection highlights that areas with significant initial development are less prone to further expansion, aligning with urban saturation theories.
The variable Capital exhibits a strong, negative, and statistically significant effect across all models, signifying the constrained growth potential of heavily urbanized capital cities. This aligns with urban economic theory, suggesting that capital cities, typically the most developed urban centers, demonstrate limited growth potential due to current dense development and regulatory limitations. This phenomenon highlights the tendency of highly developed urban areas to undergo slower surface development, a trend that may indicate both physical limitations and governmental limits.
The existence of a seaport indicates a significant and beneficial effect on the proliferation of artificial surfaces across many models. This indicates that areas with marine ports are expected to have substantial expansion, likely fueled by commercial activity and development linked to port infrastructure. The variable Airport demonstrates a positive and substantial effect in specific models, indicating that proximity to an airport correlates with the expansion of artificial surfaces, presumably due to improved accessibility and the economic growth it facilitates.
The distance variables included in Models 2 to 5 demonstrate substantial geographical relationships. Proximity to significant transportation and logistical networks, including railway stations, highways, airports, and ports, is associated with heightened artificial surface expansion, highlighting the attractiveness of accessible areas for urban and economic development. The coefficients for proximity to the nearest railway station, highway, airport, and harbor are all negative and statistically significant when analyzed individually, suggesting that areas nearer to these services generally exhibit elevated development rates in artificial surfaces. This corresponds with expectations, as regions with superior infrastructural connections are typically more appealing for additional development, whereas remote areas exhibit diminished potential for urban expansion.
The work utilizes rigorous statistical methods to tackle heteroscedasticity and multicollinearity, so ensuring the trustworthiness of the findings and reducing any biases. The findings enhance comprehension of the influence of transport infrastructure on land use dynamics, with ramifications for sustainable urban design and growth management in densely populated regions.
Conclusion
This study offers significant insights into the processes of urban expansion in Catalonia, emphasizing the divergent growth trends between capital cities and outlying towns. Capital cities like Barcelona, characterized by intricate planning systems, face stagnation from urban saturation, but adjacent municipalities exhibit enhanced adaptability and growth potential owing to more straightforward planning processes and higher flexibility. From 2006 to 2018, land use alterations near transport infrastructure rose by as much as 4.4%, whilst isolated regions experienced changes of less than 1%, and densely populated areas such as Barcelona exhibited growth rates below 0.5%. The analysis indicates that proximity to essential transport hubs, especially seaports and airports, significantly influences artificial surface expansion, highlighting accessibility as a fundamental element in urban development.
These findings highlight the necessity for strategic infrastructure planning to foster equitable urban development throughout Catalonia. Transport infrastructure might be linked to development in sparsely inhabited areas, facilitating the redistribution of growth pressures from congested metropolitan centres to more flexible regions. A prospective research line might include splitting the examination of artificial surface expansion (TAS) by various land-use categories to figure out if these alterations are predominantly residential, commercial, or industrial.
References
Kasraian, D., Maat, K., Stead, D., & Van Wee, B. (2016). Long-term impacts of transport infrastructure networks on land-use change: an international review of empirical studies. Transport reviews, 36(6), 772-792.
Litman, T. (1995). Evaluating transportation land use impacts. World Transport Policy & Practice, 1(4), 9-16.
(1) Total Artificial Surface (TAS): A measure encompassing all artificial areas, including urban fabric, industrial and commercial zones, mines and construction sites, as well as non-agricultural vegetated areas.
(2) GIS: Geographic Information System.
(3) EEA SDI: The European Environment Agency's Spatial Data Infrastructure.
(4) 'Near' operation: Calculates the shortest distance between features in GIS.
(5) 'Feature to Point': Converts geographic features to point locations in GIS.
(6) Geodesic: refers to accurate calculations that account for the Earth's curvature.
(7) Spatial Join: Combines attributes from one layer to another based on their spatial relationship in GIS.
(8) ETRS 1989 LAEA: A projection system for accurate spatial analysis across Europe, minimizing distortions in area and distance.
(9) 'Select by Location': A GIS tool that selects features based on their spatial relationship to other features
Document information
Published on 02/03/25
Submitted on 20/11/24
Volume Infraestructures i gestió sostenible del territori, 2025
Licence: CC BY-NC-SA license
