In Press (Autumn)
Back to the articles list |
Back to browse issues page
University of Hormozgan, Faculty of Engineering, Bandar Abbas, Iran , amirii@hormozgan.ac.ir
Abstract: (28 Views)
1- Introduction
Rapid and unplanned urban expansion has led to fragmented landscapes, reduced arable land, environmental degradation, and increased urban poverty. The United Nations estimates that by 2050, 60% of the world’s rural population will migrate to urban areas, especially in developing countries like Iran. To achieve sustainable, resilient, and inclusive cities by 2030, accurate monitoring and understanding of urban growth is crucial. Urban models—such as Markov chains, Cellular Automata (CA), and machine learning techniques—have become essential tools for simulating land use changes. While the Markov model predicts land use transitions over time, it lacks spatial awareness. Conversely, CA models incorporate spatial rules but fall short in capturing external influences. Combining these approaches in CA-Markov models has shown promise in simulating urban dynamics, especially in coastal regions facing land scarcity, sea-level rise, and topographic constraints.
Coastal areas are of strategic and economic importance, yet face limited space for urban expansion due to geographic and environmental constraints. These challenges demand predictive modeling to support better land use planning. CA-Markov models have been widely applied to simulate urban expansion in regions like Mazandaran, Chabahar, Asaluyeh, and Ahvaz in Iran, as well as international locations such as Bangladesh, China, Malaysia, and Sri Lanka. These studies demonstrate that rapid coastal urbanization often comes at the expense of agricultural land and natural ecosystems. Addressing this issue requires spatial modeling tools to forecast urban growth, evaluate planning policies, and mitigate the adverse impacts on land resources.
2- Study Area and Modeling Methods
The study area is the city of Bandar Abbas, located in the southernmost part of Iran, serving as the capital of Hormozgan Province (Figure 1). The city spans an elevation range of approximately 0.6 to 40 meters, with low-lying areas near the coast and higher elevations in the north. Bandar Abbas has experienced rapid urban growth, with its population rising from 17,700 in 1956 to over 680,000 in 2016. Due to natural and man-made constraints such as mountains, industrial zones, railways, and the coastline, the city’s physical expansion faces several limitations. This study aims to assess the spatial distribution and evolution of Bandar Abbas's urban form over recent decades and predict its future development trajectory using spatial modeling tools, helping determine whether the city's growth is proceeding sustainably.
To analyze land use/land cover (LULC) changes and predict future urban expansion, Landsat satellite imagery (TM, ETM+, and OLI) from 1990 to 2020 was processed using the Maximum Likelihood Classification (MLC) algorithm in ENVI software. Change detection was conducted using a post-classification comparison method and the Land Change Modeler (LCM) with CA-Markov chains in the Idrisi software. To enhance classification accuracy, preprocessing involved radiometric and geometric corrections. A 25 × 15 km region of interest (ROI) was selected for analysis. Conversion potential for urban growth was modeled based on variables such as proximity to roads, fault lines, railways, watercourses, rangelands, industrial zones, and flood-prone areas, using fuzzy logic and Cramer's correlation coefficient. An artificial neural network within LCM was employed to simulate and forecast land use dynamics.
3- Results
In this study, satellite imagery from Landsat for the years 1990, 2000, 2013, and 2020 was classified using the Maximum Likelihood algorithm into three main land use classes: water bodies, barren lands, and urban-settlement/green space areas. Due to the dry climate and arid environment of Bandar Abbas, vegetation is sparse and mostly consists of saline desert shrubs, which were grouped under barren lands due to the 30-meter resolution of the Landsat images. Urban green spaces such as neighborhood parks were classified together with residential and commercial zones, while water bodies included the Persian Gulf, fish farming ponds, and internal water areas. The area of each class was calculated, revealing trends over 30 years.
Between 1990 and 2020, barren lands showed a decreasing trend, dropping from 25,990 hectares to 21,449 hectares, with the steepest decline occurring between 2000 and 2013. Water bodies remained relatively stable, except for a dip in 2000, followed by a rise due to fish farms and sea-level changes. Urban and green space areas steadily increased, particularly from 2000 to 2013, reflecting urban expansion in Bandar Abbas. Residential areas expanded significantly in the eastern and northeastern parts of the city, especially near the airport and main roads. Major neighborhoods such as Golshahr, Tohid, and Elahiyeh saw dense construction, while western areas like Shahid Bahonar and Payambar Azam neighborhoods also developed.
Land use maps were validated using ground control points and false color images, with Kappa coefficients above 0.7, indicating reliable classification. The highest accuracy was achieved with the OLI sensor (15m resolution), outperforming older TM and ETM sensors. Land conversion analysis showed that the most significant change was the transformation of barren lands into residential and urban spaces, marking the physical growth of Bandar Abbas. The results provide critical insights for urban planning and environmental monitoring in arid coastal cities.
4- Discussion & Conclusions
This study utilized Landsat satellite imagery (TM 1990, ETM 2000, and OLI 2013 and 2020) to analyze land use/land cover (LULC) changes in a 25×15 km area in Bandar Abbas. Through classification using the Maximum Likelihood algorithm and subsequent change detection with LCM in Idrisi TerrSet, significant urban growth trends were identified. The CA-Markov model was applied to forecast urban development in 2050, revealing that the most substantial urban expansion occurred between 1990–2000 and 2000–2013. Urban areas were classified into three main categories—water bodies, urban-residential with green spaces, and barren lands—due to the arid climate and sparse vegetation. Most green spaces are artificial and merged with residential zones due to resolution constraints.
The results show that approximately 4,480 hectares of barren land were converted to residential use over 30 years, particularly in eastern and northeastern neighborhoods, coastal commercial zones, and the newly developed northwest areas. Natural barriers like the sea to the south and mountainous terrain to the north have constrained city expansion, directing urban growth toward the east and northeast. Compared to other Iranian cities like Ahvaz and coastal cities along the Caspian and Persian Gulf, Bandar Abbas shares similar urbanization patterns influenced by natural constraints, migration, and industrial growth. However, unlike cities surrounded by fertile farmland, Bandar Abbas’ surrounding barren and saline lands make agricultural land-use change less relevant, emphasizing urban development over rural-to-urban transitions.
Keywords: Remote Sensing (RS), Geographic Information System (GIS), Cramer's V Coefficient, Coastal Urbanization, Markov Chain.
Received: 2025/04/24
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
