Modeling Production-Living-Ecological Space for Chengdu, China: An Analytical Framework Based on Machine Learning...

Area of Science:

• Urban planning and environmental science
• Geographic Information Systems (GIS) and spatial modeling
• Sustainable development and land use management

Background:

• Growing urban populations and land expansion intensify conflicts within productive-living-ecological spaces (PLES).
• Accurate dynamic assessment of PLES indicators is crucial for effective land use simulation and planning.
• Current models lack complete coupling between urban system evolution and PLES utilization schemes.

Purpose of the Study:

• To develop a novel scenario simulation framework for urban PLES development.
• To dynamically couple key environmental factors with PLES utilization configurations.
• To provide a tool for understanding and managing complex land space changes in urban environments.

Main Methods:

• Developed a dynamic coupling model integrating Bagging-Cellular Automata (Bagging-CA).
• Employed automatic parameterization for adjusting driving factor weights across different scenarios.
• Utilized machine learning and multi-objective scenarios with fine land use classification data.

Main Results:

• Generated diverse environmental element configuration patterns for urban PLES.
• Successfully applied the framework to case studies in Southwest China.
• Demonstrated the automatic parameterization of environmental elements for informed decision-making.

Conclusions:

• The developed framework offers a robust method for simulating PLES evolution under various scenarios.
• Automatic parameterization enhances understanding of land space changes, aiding policy formulation.
• The multi-scenario simulation approach provides valuable insights applicable to PLES modeling globally.