Forecasting Rice Field Conversion in Boyolali Regency Using CA-ANN and Multitemporal Remote Sensing Data

¹ Faculty of Geography, Muhammadiyah University Surakarta, Surakarta 57162, Central Java, Indonesia
² Faculty of Geographic Information Science, Muhammadiyah University Surakarta, Surakarta 57162, Central Java, Indonesia

^* Corresponding author: dns104@ums.ac.id

Abstract

Land conversion or land use change is a change in the function of land to another use caused by a change in the value of the land. Changes in land use occur due to various factors, such as population growth, urbanization, and infrastructure expansion, which cause rice fields to be converted into non-rice fields. In recent years, Boyolali Regency has experienced a decline in the area of rice fields due to land use change. This study aims to analyze the change in the area of rice field land use to non-rice field land and predict changes over the next 10 years. Cellular Automata - Artificial Neural Network (CA-ANN) is the method used in this analysis. Additionally, to analyze and predict changes in land use, driving factors such as distance from roads, distance from rivers, and distance from settlements are required. The results of land use changes in 2014 and 2024 show an increase in non-rice field land use of 9,846.29 ha (16.15%), while rice field land use decreased by 9,846.9 ha (20.26%), indicating substantial land changes. The results of the prediction modeling showed a significant increase in non-rice field land use of 72,774.03 ha, while rice field land use decreased by 36,818.45 ha. Modeling with CA-ANN showed strong accuracy results, with a Kappa index of 0.73350 and a correction of 87.94264%.