Approach maximum likelihood classification and cellular automata markov chain model for land use/land cover change prediction in Nagan Raya Country, Indonesia

Abstract Forest is one of the land cover classes found in Nagan Raya District. Deforestation can contribute to a decrease in carbon emission absorption potential. Land cover change can affect a region’s policy in managing a good environment. This study aims to determine land cover change for the period 2007-2023 and predict land cover in 2035. Land cover change analysis uses Maximum Likelihood Classification (MLC) and Cellular Automata Markov Chain to predict land cover in 2035. Land cover was classified into 8 (eight) groups using kappa coefficient estimation. The kappa result for land cover in 2007 was 77%. The kappa result for land cover in 2015 was 84% and the kappa result for land cover in 2023 was 82%. The dominating land cover in 2023 is forest with the area of 169.555,44 Ha. Forest deforestation in 2007-2023 amounted to 19.926 Ha (10,5%) and the addition of farmfield is 9.428 Ha (9,4%). The validation of the 2035 model is 0,79, in which the declining land cover classes are forests 12.848 Ha (7,6%), wetlands 1.518 Ha (10,8%), rice field 131,2 Ha (2%), open land 4.935 Ha (38,1%) and water bodies 164,2 Ha (4,2%). While the increasing land cover classes are settlement 1.502 Ha (20,1%) and farmfield 20.709,9 Ha (18,9%). The results of this study can support policy decision-making and regional planning systems in Nagan Raya District.