Land Use Changes Prediction in the Phra Phimon Irrigation Scheme Using Multi-Temporal Landsat Imagery and CA-ANN Model

This study analyzes land use changes in the Phra Phimon Operation and Maintenance Project, a large irrigation scheme near Bangkok, Thailand, where rapid urbanization has led to significant modifications in land use and agricultural activities. Landsat satellite imagery from 1993, 2006, and 2019 was used to classify land use into four categories: paddy fields, orchards, urban areas, and water bodies. Future land use changes for 2032 and 2045 were predicted using the CA-Markov coupled with the ANN model (CA-ANN), implemented through QGIS's MOLUSCE plugin. Land use classification achieved 90.95% accuracy (F1-score: 0.91), and prediction accuracy was 67.84% (F1-score: 0.67). Results show significant urban expansion, with urban areas projected to increase from 23.65% in 2019 to 34.87% in 2045, while agricultural areas decrease from 73.28% to 62.00%. This urban growth may result in agricultural land fragmentation, decreased water distribution efficiency, and potential irrigation canal deterioration. This research provides insights for irrigation scheme managers to prepare for future water requirements and infrastructure management changes.