Prediction of Urban Expansion and Identifying Its Impacts on the Degradation of Agricultural Land

Abstract

Over the past few decades, rapid urban development has created significant impacts on land-use/land-cover (LULC) parameters (water bodies, built-up areas, agricultural and bare land). According to the annual report of the Rajshahi district, almost 10% of the total agricultural land has been lost in the past 20 years, due to the tremendous pressure for development and subsequent sizeable LULC changes. Monitoring and prediction of urban growth and LULC change scenarios are some fundamental approaches for conserving agricultural land and ensuring sustainable urbanization and food security. This study aims to identify the loss of agricultural land using the support vector machine classification algorithm for the years 1999, 2009 and 2019, using Landsat (TM/OLI) satellite images in the Rajshahi district. The study also predicts the future LULC change scenario, which includes the estimation of urban expansion and loss of agricultural land for the year 2029, using the machine learning-based cellular automata (CA) approach. The accuracy of the LULC prediction CA model was more than 85%, based on the validation result. The result suggests a significant increase in the built-up area (+11.71%) and a decrease in agricultural land (−12.40%) in the study area. The prediction indicates that the built-up area will increase by 15.06%, whereas 9.77% of agricultural land will be converted to other land-uses through urban expansion by 2029. The significant loss in agricultural land destabilizes biodiversity and ecosystem services and increases food insecurity. The city authorities need to implement effective strategies with the help of urban planners, environmental engineers and policymakers to conserve the agricultural as well as natural wealth, which will ensure sustainable, planned and inclusive urban development in the study area in the future.