Deriving suitability factors for CA-Markov land use simulation model based on local historical data

Area of Science:

• Geographic Information Systems (GIS)
• Spatial Analysis
• Environmental Modeling

Background:

• Multiple Criteria Evaluation (MCE) is crucial for land suitability analysis and land use simulation.
• Existing MCE methods often lack quantitative factor selection, scoring, and weighting based on local historical data.
• Cellular Automata (CA)-Markov models are widely used for land use change simulation.

Purpose of the Study:

• To explore the feasibility of quantitatively selecting, scoring, and weighting MCE factors using historical land use data.
• To develop and validate an updated MCE method for improved land use simulation.
• To integrate the updated MCE with the CA-Markov model for enhanced spatial prediction.

Main Methods:

• Developed logistic regression models fitted by historical land use changes to select and score potential factors.
• Employed the Entropy method to quantitatively determine weights for selected MCE factors.
• Utilized the MCE output as input for the CA-Markov model to simulate land use changes from 2001 to 2011.
• Validated simulation results against observed 2011 land use data.

Main Results:

• The MCE factors derived from historical data demonstrated a reasonable goodness of fit for land use simulation.
• The updated MCE method successfully integrated quantitative factor selection, scoring, and weighting.
• The CA-Markov model, using the updated MCE, produced reliable land use change simulations.
• The approach allows for efficient calibration of the CA-Markov model.

Conclusions:

• The quantitative MCE method, utilizing historical data, offers a robust approach for land use simulation.
• Deriving factors, scores, and weights from local historical trends enhances the accuracy and relevance of MCE.
• This quantitative approach facilitates efficient CA-Markov model calibration and the development of land use planning scenarios.