Reverse engineered flood hazard mapping in Afghanistan: A parsimonious flood map model for developing countries

Abstract

Most flood models are based on advanced algorithmic and multiple data requirements that are sometimes difficult to apply in developing countries. These feed-forward models cannot be applied to large areas and can lead to extreme over/under estimations in some developing countries due to extrapolation from inadequate datasets where each additional parameter adds further uncertainty. This study proposes to employ a parsimonious model that only relies on adequate available data reducing forward-uncertainty-propagation. A “reverse engineering” approach that relies on past inundation depths does provide a solution for flood hazard mapping where extracting the flood extent of extreme floods is the primary goal and where only inadequate hydrological input data are available. The feedback method was successfully deployed to create the nationwide Afghanistan Flood Hazard Map (AFG-FHM) at a scale of 1:100,000 using a high-resolution digital elevation model, sample measurements and Dartmouth Flood Observatory past flood data. This paper describes the parsimonious flood map model and general methodology employed to create the AFG-FHM, as it is a robust method to generate extreme inundation outlines, which can be utilised in other developing nations as well.