Chapter 17 - Prediction of land cover changes of Khagrachhari Hilly Upazila using artificial-neural-network-based cellular automata model

Abstract

The hilly area’s land use ensures ecosystem and environmental balance, but changing these land uses may disrupt the entire environment. The Artificial-Neural-Network-based Cellular Automata (ANN-CA) Model is used in this study to estimate land-use changes in Khagrachhari Sadar, one of the extensively hilly Upazila’s of Chittagong Hill Tracts (CHTs) of Bangladesh from 1990 to 2014, as well as to simulate changes for 2020 and make predictions for 2050. The findings revealed a gradual increase in vegetation coverage from 1990 to 2020. Almost 24.24% of vegetation coverage has increased in the last 30 years, with a 7.95% decrease predicted by 2050. On the other hand, barren areas decreased about 24.47% from 1990 to 2020 and are projected to be increased by 8.83% in the next 30 years. The real map of 2020 exposed that nearly 6033.42ha area has been converted to vegetation coverage since 1990 to the present, and it is predicted that the forest coverage will decrease by 2069.55ha in the next 30 years if other situations are constant. When the projection is done using real map 2020 as input, the Kappa index shows that the overall Kappa accuracy for the prediction was 85.82% and the Kappa value was 0.72. This indicates a strong agreement. This suggests that the model used was an adequate method for simulating changes in land use and land cover, and that the method can be used to analyze land-use changes in other areas.