Erosion vulnerable area assessment of Jamuna River system in Bangladesh using a multi-criteria-based geospatial fuzzy expert system and remote sensing

Abstract

Jamuna, a dynamic and unstable braided river system in Bangladesh, is approximately 240 km long and becomes extremely unstable during the rainy season resulting in serious bank erosion. Therefore, this study aimed to assess the erosion-prone areas adjacent to the Jamuna River system. Change detection analysis was carried out using Landsat 8 (OLI) images captured in 2020 by multi-criteria analysis using a geospatial fuzzy expert system and state-of-the-art remote sensing technology. Normalized difference vegetation index (NDVI), soil-adjusted vegetation index (SAVI), normalized difference water index (NDWI), distance from the river, land use and land cover (LULC), and slope and elevation were selected as criteria for this analysis. All criteria maps were standardized using fuzzy membership functions and reclassification of each criteria performed. Furthermore, expert judgments were included to rank the criteria influencing vulnerable areas based on an analytical hierarchy process (AHP) approach. Finally, a weighted overlay map was prepared for erosion vulnerability assessment from the reclassified maps. From these analyses, we found that water bodies covered 1003 km² (10.94%), high-to-moderate erosion-prone areas were 7401.21 km² (77.39%), marginal erosion-prone areas 1065 km² (11.61%) and nonerosion-prone areas only 5.9 km² (0.06%), respectively. To verify the vulnerable areas, 150 reference points of water bodies from the mainstream of the Jamuna River were taken using Google Earth Pro images captured in 2020. These points were plotted on the NDWI maps of 2020 and 1990 to verify the detection of changes in the riverbank shifts for 30-year intervals. This confirmed the bank shifted from 3 to 4 km in more than 20 points during this span of time. Our analysis also confirmed that high-to-moderately erosion-vulnerable areas fall between 3 and 7 km. Therefore, we recommend the adoption of new agricultural land use planning, considering erosion venerable areas to ensure agricultural production and livelihood security.