International Journal of Engineering and Information Systems (IJEAIS)

Title: Modelling Flood-Driven Soil Erosion in Toru-Orua, Bayelsa State (2017-2024)

Authors: Gbenizibe Bonus Wombu, Raymond Alex Alex Ekemube

Volume: 9

Issue: 6

Pages: 70-83

Publication Date: 2025/06/28

Abstract:
: Flood-induced soil erosion poses significant environmental challenges in the Niger Delta's riverine communities. This study develops a theoretical modeling framework to estimate soil erosion in Toru-Orua (Sagbama LGA, Bayelsa State, Nigeria) from 2017 to 2024, using open-source data and parallel modeling approaches. Empirical point measurements and field data inform two complementary models: the data-light Universal Soil Loss Equation (USLE) and the physics-rich Soil and Water Assessment Tool (SWAT). Rainfall intensity data from the Nigerian Meteorological Agency (NiMet), river discharge records from the Nigerian Hydrological Services Agency (NIHSA), soil properties from FAO's Harmonized World Soil Database, Sentinel-1 Synthetic Aperture Radar (SAR) flood extents, and a 30 m Digital Elevation Model (SRTM) from NASA jointly serve as inputs. Comparative analysis includes Amassoma and Odoni, two hydrologically similar Niger Delta communities, to evaluate model transferability. The methodology integrates QGIS for spatial data processing and Python (geopandas/rasterio) for analysis automation, ensuring reproducibility and data transparency. Results indicate that USLE's long-term average erosion estimates provide a baseline but may underestimate flood-driven spikes, whereas SWAT captures dynamic runoff-sediment processes with finer temporal resolution. A comparative performance assessment shows both models identifying high-risk erosion periods during extreme floods (e.g. 2018 and 2022), with SWAT predicting slightly higher annual soil loss in all communities due to its inclusion of event-driven sediment peaks. Three visuals support the findings: (1) a study area map situating Toru-Orua, Amassoma, and Odoni in the Niger Delta floodplain; (2) a schematic workflow illustrating data inputs and modeling steps for USLE and SWAT; and (3) a results chart comparing annual soil loss predictions by model and site. The discussion addresses model assumptions - USLE's empiricism vs. SWAT's process complexity - and their implications for real-world generalization. While no policy prescriptions are made, the study underscores the importance of model selection on interpreting erosion risks under flood conditions. We conclude that a hybrid approach, leveraging USLE for broad-scale screening and SWAT for detailed scenario analysis, can enhance understanding of flood-driven erosion in data-sparse regions. Theoretical rigor and empirical validation are emphasized to improve confidence in model outputs for similar flood-prone landscapes. Acknowledgments highlight the open-data initiatives and collaborative efforts that enabled this research. The work contributes to the academic discourse on erosion modeling by transparently comparing model frameworks and exploring their applicability in a changing hydro-climatic context.

Download Full Article (PDF)