New possibilities in extreme flow event modelling and evaluation

Abstract:

Hydrodynamics models are a key element in any flood risk analysis system by providing distributed predictions of water depth, inundation extent and flow velocity. The accuracy of flood hazard maps suffers from a lack of distributed observations of inundation outlines. Unfortunately the collection of relevant aerial or satellite images providing the required spatial data is still a challenge because of the high costs. Moreover, the temporal resolution of satellite maps is often too low, and the process of obtaining images strongly depends on the weather conditions.

Remote sensing based on unmanned aerial vehicles (UAV) creates new opportunities for the acquisition of spatial data describing riverine topography. The essential advantage of UAV-based remote sensing is its effectiveness in capturing ultra-fine resolution imagery and to collect data in a flexible time frame, practically on demand, when necessary, with adequate temporal discretization. Rapidly decreasing costs of using UAV technology make it an increasingly popular data acquisition technique. In hydrological studies UAV data can provide precise information on the location of a river’s shorelines, channel geometry and vegetation. This information can be used for the calibration and validation of distributed flow routing models. The improvement of distributed model predictions in flood modelling is still a challenging problem that should be studied to increase the reliability of flood risk assessment. Due to rapid environmental changes caused by natural and human intervention, river landscapes undergo systematic transformations. The non-stationarity of river reach boundary conditions limits the real-time predictive capability of existing models. To increase their utility, models need to be updated more frequently. Other applications include the prediction of the effects of extreme hydrological events on aquatic ecosystems and investigations of sediment behaviour at high and low water levels. Our overall goal is to present a new procedure for hydrodynamic model conditioning. The approach focuses on ongoing updating of boundary conditions of the hydrodynamic model (based on UAV-born data) and is also a very effective means of on-line updating of flood risk maps and their verification.