Abstract
Most urban areas in the Middle East and North Africa (MENA) region have experienced unprecedented growth rates over the past few decades to absorb population increase and hefty migration from surrounding rural and/or politically unstable areas. Such expedited urbanization presents substantial stresses to these urban areas' ecological and financial resources, as well as to the overall well-being of their residents. This paper studies the urban growth pattern in five major cities in the MENA region including Dubai (United Arab Emirates (UAE)), Cairo (Egypt), Doha (Qatar), Casablanca (Morocco), and Riyadh (Kingdom of Saudi Arabia (KSA)). The study adopts a machine learning (ML)-based modeling framework, which integrates remote sensing and computer vision technologies to generate high-fidelity urban growth prediction with limited data requirements. The framework treats successive satellite images for the urban area under study as a video for which a future frame is constructed to present the predicted growth in a specific target year. The methodology is shown to produce growth prediction results that are consistent with previous studies conducted for these five cities. The obtained results are used to derive several recommendations to assist in developing sustainable growth policies for these cities.
•Studying the urban growth pattern for five major cities in the MENA region•Machine learning model that integrates remote sensing and computer vision technologies•Model validation using the results of previous studies conducted for the region•Providing urban growth prediction capabilities for urban planners and policy makers•Deriving recommendations to develop sustainable growth policies for MENA cities and other cities around the world