Improving land-use change modeling by integrating ANN with Cellular Automata-Markov Chain model

Area of Science:

• Environmental Science
• Urban Planning
• Geographic Information Systems (GIS)

Background:

• Urban growth and land-use change present complex challenges for future city planning.
• Accurate temporal and spatial modeling is crucial for simulating future land-use changes.
• Artificial Intelligence (AI) is increasingly used to predict urban expansion patterns.

Purpose of the Study:

• To enhance the simulation capability of the Cellular Automata Markov Chain (CA-MC) model for predicting land-use changes.
• To integrate Artificial Neural Networks (ANN) into the CA-MC model to incorporate key driving forces of land-use change.
• To improve the accuracy of urban expansion predictions for better land management strategies.

Main Methods:

• Integration of Artificial Neural Networks (ANN) with the Cellular Automata Markov Chain (CA-MC) model.
• Utilized socio-economic, spatial, and environmental variables (slope, distance to road, urban centers, commercial areas, density, elevation, land fertility) to generate transition maps via an ANN data-driven model.
• Calibrated and validated the enhanced CA-MC model against the original CA-MC model using 2015 land-use data for Irbid city, Jordan, employing Kappa indices for comparison.

Main Results:

• The integrated ANN-CA-MC model achieved a prediction accuracy of 90.04%, outperforming the original CA-MC model's accuracy of 86.29%.
• The study demonstrated that incorporating driving forces through ANN significantly enhances prediction accuracy for land-use change.
• Validation confirmed superior performance in both the quantity and location of predicted land-use changes.

Conclusions:

• The integration of ANN with CA-MC provides a more accurate prediction of urban land-use change by accounting for influential driving forces.
• The enhanced model offers valuable tools for urban planners and local authorities to develop sustainable management strategies.
• Improved land-use predictions support balancing urban expansion with agricultural protection, crucial for regional food security.