Prediction of urban expansion by using land cover change detection approach

Abstract

Bangladesh has been experiencing rapid urban expansion over the last few decades, contributing much to the region's land cover transition into the urban area. The study aims to employ geospatial modeling techniques to investigate land cover scenarios in the Pabna municipality of Bangladesh. Therefore, the research examined Cellular Automata Markov and Multi-Layer Perceptron Markov models to detect land cover for 2023 and 2028. The study selected the Multi-Layer Perceptron Markov as the best fit model over Cellular Automata Markov based on the highest kappa value. The result reveals that urban area has increased from 3.39 to 8.79 km2 over 1998–2018. Urban expansion and its surrounding area are primarily occurring towards the northeast directions. However, the extent of urban build-up land will grow from 3.39 km2 in 1998 to 11.01 km2 in 2023 and 12.44 km2 in 2028. Moreover, the future land cover map delineated that the urban growth will expand in the northeast part of the study area. The scenario shown in this paper would assist urban planners in quantifying the urban growth under different land cover features and thus preparing proper strategic measures.

Keywords

• Urban expansion
• Land cover
• GIS
• Cellular automata
• Multi-layer Perceptron Neural Network