The Impacts of Flood on the Agricultural Sector in Chalan Beel Area: A Study on Tarash Upazila

Flood-induced agricultural disruption is a growing hydro-climatic threat in the low-lying monsoon floodplains of Bangladesh, where riverine inundation, wetland hydrology, rainfall variability, and intensive rice cultivation affect food security and rural livelihoods. Tarash Upazila of Sirajganj District, within the Chalan Beel wetland–agricultural system and Atrai–Baral floodplain, is highly exposed to seasonal flooding. However, spatially explicit evidence linking flood extent, inundation duration, elevation, crop-calendar timing, and vegetation recovery to agricultural disruption remains limited. This study assessed flood-induced agricultural impacts in Tarash from 2018 to 2023 using an integrated multi-sensor remote sensing and GIS framework in Google Earth Engine.

Sentinel-1 C-band SAR imagery (Sentinel-1A/B where available) was used for flood mapping because of its all-weather capability. Annual flood extent and duration were derived using VV/VH backscatter change detection, speckle filtering, and automated Otsu thresholding. Flood classifications were validated through confusion-matrix assessment using temporally matched Sentinel-2 reference classifications and georeferenced field observations, yielding 89.2% overall accuracy and flooded-class producer and user accuracies of 86.7% and 91.5%, respectively. Sentinel-2 MSI and Landsat OLI imagery were used to map agricultural land and assess vegetation and surface-water conditions using NDVI, EVI, LSWI, NDWI, and MNDWI. MODIS NDVI time series, DEM-derived elevation, GPM rainfall, and Aus, Aman, and Boro crop calendars were integrated to evaluate crop-specific exposure, vegetation stress, and recovery. Agricultural land estimates were cross-checked against Bangladesh Bureau of Statistics (BBS) agricultural statistics.

Agricultural land accounted for approximately 62–68% of Tarash during 2018–2023. Annual monsoon flood exposure varied substantially, affecting a minimum of 38% of mapped agricultural land in 2018 and a maximum of 46% in 2020, with no consistent temporal trend. Prolonged inundation (>15 days) affected 21–27% of total agricultural land, predominantly in low-lying central, western, and southwestern areas. Aman was the most exposed rice crop, accounting for 57–64% of flood-affected cropland. Mean NDVI reduction in flooded croplands increased from 0.16 in 2018 to 0.24 during the more extensive 2020 flood, while recovery toward pre-flood NDVI conditions increased from approximately three to five weeks, respectively.

These findings demonstrate that agricultural disruption is influenced by flood extent, inundation persistence, crop-calendar timing, and vegetation recovery dynamics. The framework provides a transferable approach for crop-specific flood-impact assessment and flood-resilient agricultural planning in vulnerable wetland regions of Bangladesh.