LULC and Drainage Dynamic Study of Polders in Bangladesh

Abstract

The southwestern part of the Ganges-Brahmaputra coastal area is facing a severe waterlogging problem due to both natural and human alteration of the landscape while affecting the lives and livelihood of people. The expansion of permanently inundated lands in the form of shrimp farming is one of the major challenges to the socio-economic development of the southwestern region. The primary objective of this study is to present a methodology that can evaluate the critical reason behind the increasing shrimp farming in the southwest polders of Bangladesh. This study investigates temporal changes in land use and land cover (LULC) patterns within the polders of the southwestern region. Landsat images of the last 48 years have been studied through image classification to detect the transformation of areas with an interval of mostly ten years. Three dominant land use classes presented in the study area, such as Water, Vegetation, and Lands and Settlement, were considered when comparing the Land use and Land cover change. The results revealed that the waterlogged area has increased during the study period, with a significant change in the later ’90s. At the same time, a decreasing trend in the land and settlement category has been seen. However, the vegetation areas (exempting the agricultural land) have been increased over the years. A post-classification comparison between NDVI images with the classified images was performed to validate the classified images. Rice cultivation has been declined significantly with the rapid extension of shrimp farming in the southwest coastal region of Bangladesh. However, inadequate drainage, scarcity of freshwater resources, salinity intrusion in surface water, and less workforce involvement in shrimp farming than rice cultivation might have aggravated the process. Therefore, in continuation of this study, a drainage model has been developed to evaluate if inadequate drainage has played any essential role in shifting from rice cultivation to shrimp farming in polder #15. A Rainfall-runoff (NAM)and a hydrodynamic (HD) model of the MIKE11 software have been used in developing the drainage model of polder #15. The drainage canals that drain in proportion to their respective drainage areas, the drainage regulators, and surrounding rivers of the polder catchment area are the main features of this developed drainage model. ArcGIS has been used in DEM preparation, river network preparation, and processing of the 4 catchment boundaries that are required in MIKE11 modules. However, structures included in this model are according to their existing dimensions, number of vents, and invert levels. Furthermore, the operation rules for the structure gates are defined, such as that if the outside water level is lower than that of polder water level in the drainage khals/canals, the gates will be automatically open and vice versa. Because of the scarcity of data in the coastal area, the discharge and water level time series for upstream and downstream boundaries have been collected from the existing South West Regional Model available at the Institute of Water Modelling, Dhaka. Rainfall based on a 5 days maximum cumulative rainfall event with a 10-year design period has been considered for simulation during the frequency analysis. However, 3-days water level inundation depth has been analyzed to evaluate the drainage performance of the polder. The model result shows that the polder#15 drainage system is insufficient considering the 10-year design rainfall event. Furthermore, the comparison between the seasonal inundation depth maps of 1990,1994,2000, and 2010 have indicated inadequate drainage provisions in the monsoon season, which plays a crucial role in increasing shrimp farming.