Urban growth simulation and scenario projection for the arid regions using heuristic cellular automata

Abstract

Abstract Arid regions tend to form compact urban patterns that have significant implications on urban growth and future urban patterns. Spatial simulation and projection using cellular automata (CA)-based models are important for achieving sustainable urban development in arid regions. In response to this need, we developed a new CA model (GSA-CA) using the gravitational search algorithm (GSA) to capture and project urban growth patterns in arid regions. We calibrated the GSA-CA model for the arid city of Urumqi in Northwest China from 2000 to 2010, and validated the model from 2010 to 2020, and then applied to project urban growth in 2040. The results indicated that the optimal performance of the model was achieved when the fraction of the population was 0.5. GSA-CA achieved an overall accuracy of 98.42% and a figure of merit (FOM) of 43.03% for the year 2010, and an overall accuracy of 98.52% with FOM of 37.64% for 2020. The results of the study help to adjust urban planning and development policies. The developed model has the potential to be employed in simulating urban growth and future scenarios in arid regions globally, including Northwest China and Africa.

Keywords

• Gravitational search algorithm
• Arid regions
• Urban growth
• Cellular automata
• Scenario projection