Modeling and Forecasting Urban Sprawl in Sylhet Sadar Using Remote Sensing Data

Abstract

Forecasting urban sprawl is important for land-use and transport planning. The aim of this study is to model and predict the future urban sprawl in Sylhet Sadar using remote sensing data. The ordinary least square (OLS) regression model and the geographic information system (GIS) are used for modeling urban expansion. The model is calibrated for the years 2014 to 2017 using eight explanatory variables extracted from the regression model. The regression coefficients of the variables are found statistically significant at a 99% confidence level. The cellular automata (CA) model is then used to analyze, model, and simulate the land-use and land-cover (LULC) changes by incorporating the algorithm of logistic regression (LR). The calibrated model is used to predict the 2020 map, and the result shows that the predicted map and the actual map of 2020 are well agreed. By using the calibrated model, the simulated prediction map of 2035 shows an urban cell expansion of 220% between 2020 and 2035.