Flood Risk Assessment in Poorly Gauged River Basins -A Case Study of the Oti River Basin, Togo, West Africa

Abstract

In many parts of the world, hydrological extremes have been observed and grave consequences on human life, socio-economic activities and ecosystems have been reported. Over the last few years, West Africa has increasingly experienced severe floods that have affected millions of people and hampered economic development in the region. Floods in Oti River Basin result in loss of life and property, damages of growing crops and infrastructure, and disruption of social and economic activities. These consequences could be exacerbated by the effects of human-induced climate change. Therefore, managing the risk of extreme events and disasters is important for an efficient climate change adaptation. In West Africa, the management of floods becomes a strategic priority at local, national and regional levels, particularly in the aftermath of the severe floods in 2007, 2008 and 2010. Flood risk assessment is the first step of any flood risk management project. Such assessment gives insight on what can be expected and opens up the discussion on how to tackle the situation. However, flood risk information is missing in many West African catchments basically in the Oti River Basin due to the lack of the hydrological data required to estimate it. The urgency of this subject and the existing research gaps in the study area are the main motivation for the present work. This thesis is an initial detailed flood risk assessment in the Oti River Basin (Togo). The research methodology was articulated in four main steps namely (i) regional flood frequency analysis based on L-moments (ii) hydrological modelling for flood prediction, (iii) flood hazard mapping using flood inundation model and (iv) community-based flood risk assessment via a combination of flood hazard map, survey data on vulnerability, exposure and coping capacity. The results of this study showed that Generalized v Extreme Values (GEV) and Generalized Pareto (GPA) are the best probability distributions to estimate flood frequency in the Oti River Basin. For the hydrological modelling, Nash-Sutcliffe coefficients of 0.87 and 0.94 were respectively obtained for the calibration and validation periods. Moreover, the present study showed that flood extends in the Oti River Basin doesn’t change drastically for higher return period floods. This thesis showed also that flood risk in the Oti River Basin is moderate and predominantly driven by high vulnerability and hazard. Both exposure and coping capacity are relatively low. Finally, further research is needed to understand the effects of rainfall on the variability of L-moments of annual maximum discharges, to analyze the combined effects of climate and land use changes on flood risk and to implement advanced flood early warning system in the Oti River Basin of Togo.