PREDIKSI KONVERSI LAHAN PERTANIAN BERBASIS ARTIFICIAL NEURAL NETWORK-CELLULAR AUTOMATA (ANN-CA) DI KAWASAN SLEMAN BARAT

Abstract

Analysis and prediction of land conversion using spatial-temporal data are essential for environmental monitoring and better land use planning and management. The West Sleman area has the potential to experience land use changes due to anthropogenic factors. This study aimed to determine the spatial-temporal dynamics of land use change in 2012-2022 and predict future land use change using the ANN-CA model for 20 years (2022-2042). Analyzed the spatial-temporal dynamics of land use change based on land use data derived from SPOT imagery, then predicted future land use change with the ANN-CA model using the MOLUSCE plugin on QGIS Desktop 2.18.11. The simulation results showed an accuracy of 86.66% and an overall Kappa value of 83% obtained by comparing the actual data in 2022 with the simulated data on land use change in the same year. The irrigated paddy fields decreased by 6.39% (685.22 ha) due to conversion to settlements. The area of residential buildings increased by 4.65% (498.49 ha) during 2012- 2017. Predictions of land use change in 2022-2042 show that the reduction of irrigated paddy fields will continue, and the number of residential buildings tend to increase.