Understanding future land use land cover dynamics is critical for sustainable urban land use planning, particularly in rapidly growing secondary cities like Debre Birhan, Ethiopia. This study addresses the pressing challenge of unplanned urban expansion, which threatens agricultural lands, forest ecosystems, and water resources. The objective of the study was to predict future land use scenarios using a hybrid spatio-temporal model integrating Cellular Automata (CA)-Markov chain and Multi-Layer Perceptron Artificial Neural Network (MLP-ANN). Geospatial tools were employed to analyze historical land use changes (1984–2023) and simulate future scenarios for 2062. The model calibration achieved high accuracy across multiple metrics, yielding a Kappa no of 0.9490, Kappa location of 0.9638, and Kappa standard of 0.9267. Results indicate a significant decline in agricultural (11.65%) and forest lands (1.56%), while built-up and industrial areas will expand by 5.38% and 3.18%, respectively by 2062. These trends highlight accelerating urbanization at the expense of food security and ecosystem integrity. The model’s output indicates spatially-targeted planning intervention such as protecting high-risk agricultural zones, conserving forest and riparian buffers and promoting compact urban growth with transport corridors. The hybrid CA-Markov Chain and MLP-ANN approach proved effective in capturing non-linear land use dynamics, offering a robust framework for informed decision-making.
Zegeye et al. (Sun,) studied this question.