Spatio- temporal Change Dynamics in the Estuarine Island of Bangladesh

Abstract

Rapid urbanisation-induced land use and land cover (LULC) changes alter the vegetation cover and groundwater table. This study aims to identify the land use and land cover changes, the factors responsible for this shift, and the effects of the changes on Sandwip, an estuarine island in Bangladesh. Sentinel-2 and Landsat 5 images were collected, and Google Earth Engine (GEE) was used for supervised classification and analysis. Google Earth Pro and ArcGIS 10.4 software were used for creating shapefile and map preparation. Results revealed that from 1992 to 2022, there was a substantial growth of vegetation cover (164.02%) that occurred because of the increased trees outside forest (TOF). An increasing relative growth was also recorded in settlement (776.21%) and char land (0.22%). In contrast, waterbodies decreased by 46% and agricultural land by 45.67% during the time. Almost 87% of people opined that the remittance was the most important factor behind settlement and homestead vegetation cover expansion. Open water bodies were found to be the most dynamic land use type in Sandwip. The Mean modified normalised difference water index (MNDWI) varied in every decade from 1992 to 2022. It was highest (0.2428) in 1992 and reduced to 0.0839 in 2002, then increased again to 0.1343 and became lowest (0.0099) in 2022, which indicates a lowering of the groundwater table. The overall accuracy (73%-96%) and Kappa coefficient (0.60-0.94) of the classified maps revealed that the produced maps have good consistency with the reference maps. The methodological process of this study offers a low-cost, efficient technique for monitoring land use patterns and groundwater table conditions, which can guide land use planners and policymakers for sustainable management.