Machine learning based modeling for future prospects of land use land cover change in Gopalganj District, Bangladesh

During the last decade, urban growth has been increased rapidly in Gopalganj district of Bangladesh. Therefore, this study seeks to observe fluctuations in land use and land cover (LULC) in Gopalganj with its effects over the past two decades. Through Landsat 5 (TM) and 8 (OLI) data, this research focuses on Maximum Likelihood Supervised Classification (MLSC) technique for creating land-use classes of different years. Built-up areas have increased by 21.17% over the last two decades. Urban vegetation has declined in parallel with the increase in vacant land in the district from 2000 to 2010. However, in the next ten years, it has occupied about 72.76 square km of bare-land. By using Artificial Neural Network (ANN) with integrated cellular automation (CA) simulation, the model predicted that urban areas would grow by 10.88% in the central and north-western regions of the district by 2050. Urban vegetation will decrease by 4.09%, with a significant reduction in bare land and water bodies. The accuracy of the predicted LULC is 89.48% based on validation result. This prediction may help municipal and administrative authorities, urban planners to achieve a planned and sustainable future city of Gopalganj.