A Comparative study of land use dynamics in urban and peri-urban areas of Greater Beirut Agglomeration and Greater Paris Region: a geospatial approach

• Published in: Modern Cartography Series Vol. 11; pp. 89 - 115
• Main Authors: Al-Shaar, Walid, Bonin, Olivier
• Format: Book Chapter
• Language: English
• Published: 2024
• Subjects: comprehensive land use prediction; land use classification; Remote sensing technologies; semantic similarity classification; urban modeling; urban regeneration; land use classification; urban modeling; Remote sensing technologies; semantic similarity classification; comprehensive land use prediction; urban regeneration
• ISBN: 0443158320; 9780443158322
• ISSN: 1363-0814
• DOI: 10.1016/B978-0-443-15832-2.00005-8

Urban sprawl and peri-urbanization are major concerns today, because of population growth, increased propensity to live in urban areas, environmental and socioeconomic concerns, and also because of the extensive consumption of resources. External changes, such as wars or massive migrations, can have a great impact on the urbanization process. Nevertheless, natural resources, industrial and technological evolution as well as the form of political systems play a key role in the development and attractivity of a country and need to be taken into account when analyzing the land use dynamics and urban development. Monitoring and predicting urban growth and soil artificialization/imperviousness is hard for countries where data is scarce. Countries at stake (mainly developing countries) rarely have a good statistical system. Remote sensing (images and modeling algorithms) coupled with Geographic-Information-Systems (GIS) emerged within the last 40 years as a good tool for examining the urban evolution and land use dynamics since it enables to (1) classify land use and land cover (LULC) into urban areas (dense, mixed, peri-urban, etc.) and other classes and (2) predict the evolution of LULC via several modeling approaches: Cellular Automata (CA), Markov Chain Model (MCM), CAMCM, CAMCM based models (such as SLEUTH, Dyna-CLUE). More recently, CAMCM models are coupled with techniques from artificial intelligence (AI) such as the Multi-Layer-Perceptron (MLP) and Artificial Neural Network (ANN), as well as with Agent-Based-Models (ABM), etc. Two challenges are linked to the context of peri-urbanization and urban sprawl: (1) the areas of interest are mixed LULC classes (as the mix of sparse housings with green spaces, and the mix of commercial and industrial areas, etc.), where generally the confusion error is high, hence the rural-urban linkage is not presented; and (2) the urban sprawl is subject to many external factors (increase in energy prices, new mobility services, the evolution of technology). Moreover, most countries have policies and regulations to try and mitigate the urban sprawl problem. In addition to that, the temporal transition phases of LULC, including the intermediate phases, are not systematically included. This chapter reviews the main approaches to tackle these issues, via an extensive review of literature, both on the processing of raw images and on the modeling of LULC dynamics. Recent classification techniques enable to extraction of more details from the images, as well as to add semantics to the obtained classification. The evolution of sensors gives higher resolutions images that open new fields of possibility, such as 3D processing or real-time sensing. Moreover, the coupling of CAMCM with multi-criteria decision analysis allows us to better model and understand LULC evolution. This would contribute, to the assessment, monitoring, and promotion of urban sustainability indicators at regional and global scales. Results are illustrated by a case study on the Greater Beirut Agglomeration in Lebanon, where only remote sensing images are available as data sources, and on the Greater Paris region in France, where more data exist and can be used in models.