Spatial Analysis Model of Land Use Change and Prediction in Ciawi District Using Cellular Automata - Artificial Neural Network

Abstract

Land use change is an important issue that occurs massively in the Ciawi Regency area, mainly due to rapid population growth and the increasing need for residential space, public facilities, and infrastructure. The problem that arises from this phenomenon is the occurrence of intensive land use conversion, especially the increase in built-up land area which has the potential to disrupt the environmental balance. This study aims to analyze land use changes from 2003 to 2023, as well as predict land use conditions in 2033. The study covers the entire administrative area of Ciawi Regency, Bogor, with a focus on seven land use classes, including built-up land, forests, rice fields, and gardens. The methodology used includes classification of Landsat 5 and 8 images using the Random Forest algorithm through the Google Earth Engine (GEE) platform, as well as predictive modeling using the Cellular Automata – Artificial Neural Network (CA-ANN) method through the MOLUSCE plugin in QGIS. The driving variables used in the prediction include distance to road, distance to settlement, and distance to river. The results of the study show a significant increase in built-up land from 291.45 hectares (2003) to 1,262.37 hectares (2023), and is predicted to reach 1,073.07 hectares in 2033. Prediction model validation showed an overall accuracy of 92.92% and a Kappa coefficient value of 0.82, which signifies excellent model quality. These findings are expected to be the basis for consideration in spatial planning and sustainable development policies in Ciawi Regency.