Flood risk modelling using the hierarchical process analysis (HPA) method: Example of the city of Thies, Senegal

Based on a pair-wise comparison of passive factors (the physical environment (slope, permeability, porosity and granularity of soils), the natural environment (hydrography, forest and vegetation), various networks (roads, railways, drainage) and human action (land use, implementation of development policies) and intrinsic flood triggering factors, we modelled the flood risk using Hierarchical Process Analysis (HPA) in the city of Thiès (Senegal). After processing and analysing perception surveys of flood management stakeholders, we used HPA techniques to calculate coherence indices (CI) and coherence ratios (CR), which are respectively equal to 0.27 and 5% according to members of the ORSEC plan, and 0.21 and 4% according to people affected by flooding. These results show that the perception work approach is consistent and acceptable, and have made it possible to determine the flood risk index (FRI) of the stakeholders, who have many similarities. By implementing the FRI in ArcGis software, the flood risk map was generated automatically. This revealed a high risk of vulnerability of the main outlets occupied by human settlements. Our work enabled us to determine the overall vulnerability rate for the study area, which is equal to 10%, with a variant containing 4% high vulnerability affecting a total of 4,183 properties, compared with a medium vulnerability rate of 6% affecting a total of 8,847 properties.