Flood risk assessment in urban catchments using multiple regression analysis

Flood assessment in urban catchments is usually addressed through the combination of geographic information systems (GISs) and stormwater models. However, the coupled use of these tools involves a level of detail in terms of hydrological modeling that can be beyond the scope of overall flood management planning strategies. This research consists of the development of a methodology based on multiple regression analysis (MRA) to assess flood risk in urban catchments according to their morphologic characteristics and the geometrical and topological arrangement of the drainage networks into which they flow. Stormwater models were replaced by a combination of multiple linear regression (MLR), multiple nonlinear regression (MNLR), and multiple binary logistic regression (MBLR), which enabled identifying influential parameters in the maximum runoff rates generated in urban catchments, modeling the magnitude of peak flows across them, and estimating flood risk in the nodes of sewer networks, respectively. The results obtained through a real urban catchment located in Espoo, Finland, demonstrated the usefulness of the proposed methodology to provide an accurate replication of flood occurrence in urban catchments due to intense storm events favored by climate change, information that can be used to plan and design preventative drainage strategies.