The relationships between economic growth and cropland changes in Bangladesh: An evidence based on annual land cover data

Abstract

Bangladesh is experiencing rapid economic progress since 1990s, at the same time, facing acute shortage of agricultural land. Using 19 years (2000-2018) time series remote-sensing data and Random Forest machine learning algorithm in Google Earth Engine (GEE), this study classifies the land covers in Bangladesh; determines the trend of land cover changes using least-square growth rate model; and finds the associations between the trends of cropland and growth of socio-economic factors using correlation and regression analysis. The annual average area of cropland decline is 29271 hectare, which is higher than the absolute changes of other major land covers, such as forest (25932 hectare) and built-up area (1649 hectare). With the annual growth of urban population (3.81%) and gross domestic product (GDP) (6.28%), the predicted declining rate of cropland is estimated 0.29%, which is very close to the observed annual declining rate (0.28%). However, the partial effects of urban population and GDP could not be detected. Therefore, the study infers that urbanization and economic growth in Bangladesh are happening simultaneously. As a result, the joint determination of both phenomena can explain the degradation of agricultural land in Bangladesh.

Keywords

• Agricultural land
• GDP
• Google Earth Engine
• GEE
• Landsat
• Satellite imagery