Integrated model for land-use transformation analysis based on multi-layer perception neural network and agent-based...

Area of Science:

• Environmental Science
• Geospatial Analysis
• Urban Planning

Background:

• Land-use change significantly impacts aquatic ecosystems and water quality, particularly in urban settings.
• Effective water resource management necessitates understanding spatial-temporal land-use dynamics for sustainable development.
• Previous studies highlight the need for accurate land-use change prediction models.

Purpose of the Study:

• To develop and validate an integrated model combining agent-based modeling (ABM) and multi-layer perceptron (MLP) for predicting future land-use transformations.
• To assess the impact of human decision-making, specifically farmers' and urban landowners' choices, on land-use change.
• To provide a tool for evaluating the consequences of land-use changes on water resource management.

Main Methods:

• Utilized Landsat imagery from 1989, 2004, and 2019, processed on the Google Earth Engine (GEE) platform.
• Employed random forest-supervised classification for land-use mapping with high overall accuracy (0.82-0.84) and kappa coefficients (0.74-0.78).
• Integrated an agent-based model (ABM) with a multi-layer perceptron (MLP) neural network, incorporating seven driving factors to simulate land-use changes.

Main Results:

• The integrated MLP-ABM model successfully simulated land-use changes for 2019 and projected changes for 2040.
• Predicted a significant increase in residential areas (67.96 km²) and pasture lands (64.63 km²) by 2040.
• Projected a degradation of agricultural lands (84.19 km²) and barren lands (47.98 km²) between 2019 and 2040.

Conclusions:

• The developed integrated MLP-ABM model effectively predicts land-use changes driven by human decision-making.
• Findings underscore the dynamic nature of land-use patterns and their implications for water resource management.
• The model serves as a valuable tool for urban planners and environmental managers to anticipate and mitigate the effects of land-use change.