From past to future: simulating land use and land cover changes using CA-Markov and multi-temporal satellite data

Abstract

Understanding land use and land cover (LULC) dynamics is essential for sustainable environmental management and spatial planning. This study applied the Cellular Automata-Markov (CA-Markov) model integrated with multi-temporal Landsat imagery (TM, ETM + , and OLI) to analyze past changes and project future LULC transitions in the & Ccedil;ank & imath;r & imath; Forest Enterprise region of T & uuml;rkiye. The classified maps for 2000, 2010, and 2020 achieved high accuracy (overall accuracy up to 95.3%, Kappa = 0.92). The CA-Markov model, validated with a 91.5% similarity rate, effectively simulated spatial-temporal dynamics of agricultural expansion, urban growth, and forest alteration. Results reveal that agricultural land increased by 73.4% over two decades, mainly driven by demographic growth and economic incentives, while forest and water resources declined. Future projections for 2030 and 2040 indicate continued urban expansion and agricultural dominance, accompanied by shrinking water bodies and shifting forest distribution. These trends suggest potential challenges for agricultural sustainability and forest conservation under increasing resource pressures. By linking model-based predictions with national forest and land-use policies, this study underscores the importance of integrating spatial forecasting tools into adaptive planning strategies. The CA-Markov framework provides a reliable, interpretable, and transferable approach for regional-scale scenario generation and evidence-based environmental policy development.