Assessment of Flood-Induced Soil Erosion and Agricultural Yield Loss in the Kosi Basin Integrating Remote Sensing and Hydrological Modeling

Recurring monsoon floods in the Kosi River Basin pose a critical threat to agricultural productivity, sediment deposition, and disruption of cultivable flood plains in Bihar, India. Severe seasonal rainfall combined with steep Himalayan topography generates high runoff and sediment fluxes in the Kosi River system. These hydrological conditions drive frequent channel instability, abrupt shifts in the river course, and widespread flood inundation leading to extensive deposition of sandy soil over fertile agricultural lands. Such recurring flood-driven erosion and sedimentation processes necessitate a quantitative assessment of their impacts on crop productivity, land-use dynamics, and sediment redistribution. Therefore, this study aims to provide a basin-scale quantification of flood-induced soil erosion and crop productivity losses in the Kosi River Basin using integrated remote sensing and hydrological modeling approaches. The Soil and Water Assessment Tool Plus (SWAT+) was employed to simulate basin-scale hydrological processes, sediment transport, and nutrient dynamics. Model parameterization utilized high-resolution topographic data derived from the Shuttle Radar Topography Mission (SRTM) digital elevation model and land use/land cover maps generated from Sentinel-2 satellite imagery. Climatic inputs, including rainfall and temperature, were obtained from the NASA POWER climate data archive, supplemented with observed rainfall records from the Indian Meteorological Department (IMD). Observed streamflow data from the Central Water Commission (CWC), India, were used for model calibration and validation. Spatial data processing and analyses were performed using Python-based workflows within QGIS and ArcGIS environments. We also examine the historical LULC trend from satellite data to understand the spatio-temporal changes in agricultural land and floodplain. We then run future climate scenarios: bias-corrected CMIP6 projections (SSP2-4.5, SSP5-8.5) are used to drive SWAT+ simulations of future flood extent, sediment yield, and land productivity. The final results of this research activity will be presented at the Conference.

Keywords: SWAT+, soil erosion, flood modeling, Kosi River, CMIP6, LULC, sedimentation.