Modelling of flood hazard extent in data sparse areas: a case study of the Oti River basin, West Africa

Study region: Terrain and hydrological data are scarce in many African countries. The coarse spatial resolution of freely available Shuttle Radar Topographic Mission elevation data and the absence of flow gauges on flood-prone reaches, such as the Oti River studied here, make flood inundation modell...

Full description

Bibliographic Details
Main Authors: Kossi Komi, Jeffrey Neal, Mark A. Trigg, Bernd Diekkrüger
Format: Article
Language:English
Published: Elsevier 2017-04-01
Series:Journal of Hydrology: Regional Studies
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2214581817300757
Description
Summary:Study region: Terrain and hydrological data are scarce in many African countries. The coarse spatial resolution of freely available Shuttle Radar Topographic Mission elevation data and the absence of flow gauges on flood-prone reaches, such as the Oti River studied here, make flood inundation modelling challenging in West Africa. Study focus: A flood modelling approach is developed here to simulate flood extent in data scarce regions. The methodology is based on a calibrated, distributed hydrological model for the whole basin to simulate the input discharges for a hydraulic model which is used to predict the flood extent for a 140 km reach of the Oti River. New hydrological insight for the region: Good hydrological model calibration (Nash Sutcliffe coefficient: 0.87) and validation (Nash Sutcliffe coefficient: 0.94) results demonstrate that even with coarse scale (5 km) input data, it is possible to simulate the discharge along this region's rivers, and importantly with a distributed model, derive model flows at any ungauged location within basin. With a lack of surveyed channel bathymetry, modelling the flood was only possible with a parametrized sub-grid hydraulic model. Flood model fit results relative to the observed 2007 flood extent and extensive sensitivity testing shows that this fit (64%) is likely to be as good as is possible for this region, given the coarseness of the terrain digital elevation model.
ISSN:2214-5818