An Integrated Spatial Autoregressive Model for Analyzing and Simulating Urban Spatial Growth in a Garden City, China

Area of Science:

• Urban planning and geography
• Geospatial modeling
• Environmental science

Background:

• Traditional urban land-use models often lack practicality due to high precision.
• Garden city planning requires macro-level approaches beyond community-level analysis.
• Understanding spatial autocorrelation is crucial for accurate urban expansion prediction.

Purpose of the Study:

• To develop and validate an integrated spatial autoregressive (SAR), cellular automata (CA), and Markov chain model for urban land-use change prediction.
• To identify key driving forces influencing garden city expansion.
• To provide a reliable framework for urban planning and development strategies.

Main Methods:

• Construction of an integrated SAR-CA-Markov model.
• Application of the model to Chengdu, China, for urban land-use simulation.
• Validation of the model using land-use data from 2005 to 2015 (Kappa coefficient of 0.7634).

Main Results:

• The SAR model identified development trends in the southeast, optimization needs in the central region, and ecological protection for the western region of Chengdu.
• Key spatial variables influencing land use were ranked: distance to water, slope, population density, GDP, distance to main roads, distance to railways, and distance to the county center.
• The model achieved high accuracy in simulating 2015 land use, with a Kappa coefficient of 0.7634.

Conclusions:

• The integrated SAR-CA-Markov model is a reliable and accurate tool for urban land-use prediction in garden cities.
• The findings offer valuable insights for formulating effective urban space development strategies, including regional optimization and ecological conservation.
• Future land use in 2025 and 2035 can be simulated, providing a forward-looking reference for sustainable urban planning.