Prediksi Perubahan Tutupan Lahan Menggunakan Algoritma Artificial Neural Network-Cellular Automata Di Kota Medan

Abstract

The development of various sectors in urban areas is much faster compared to rural areas, making it crucial to continuously monitor land cover changes in urban regions due to their environmental, economic, and social impacts. Medan, the capital of North Sumatra Province and the third-largest city in Indonesia, experiencing growth, and expansion in that region may trigger land cover changes. This study aims to analyze land cover changes in Medan in 1993, 2003, 2013, and 2023 and predict land cover in 2033 using six driving factors: distance from roads, distance from rivers, distance from settlements, population density, elevation, and slope. This research provides valuable information on land cover changes and predictions in the city of Medan as input and consideration for the government in planning and developing the city of Medan in the future. The land cover classification used in this study follows the supervised classification method with maximum likelihood classification, while the artificial neural network-cellular automata algorithm is applied to predict land cover. The results indicate that between 1993 and 2023, the land cover category that consistently increased was settlements, reaching an area of 16,186.50 ha (57.61%) in 2023. The land cover prediction for Medan City in 2033 indicates the distribution of land cover classes as follows: cloud cover 2.97 ha (0.01%), water bodies 562.77 ha (2.00%), mangrove forests 2,174.13 ha (7.74%), mixed plantations 427.32 ha (1.52%), open land 666.27 ha (2.37%), plantations 1,401.39 ha (4.99%), settlements 17,675.91 ha (62.91%), dryland agriculture 2,205.90 ha (7.85%), rice fields 312.48 ha (1.11%), shrubs 1,065.87 ha (3.79%), and fish ponds 1,601.91 ha (5.70%).