Discussing alternative fast dike breach flood models

Dike breaches can have severe consequences<br/>due to the population centres that are often<br/>present in the hinterland. Therefore, it is vital to<br/>obtain insight in flood consequences. This<br/>knowledge can aid in setting up evacuation<br/>plans, and help design flood protection<br/>measures. Additionally, the uncertain nature of<br/>dike breach events means that authorities are<br/>interested in a range of scenarios that allow for<br/>making informed decisions.<br/>Physics-based hydrodynamic models are<br/>often the tool of choice to simulate a flood event,<br/>including those after dike breaches. The main<br/>downside is the large computational time,<br/>prohibiting scenario analysis during timesensitive emergencies. Therefore, alternative<br/>modelling techniques have been developed with<br/>the aim of decreasing computation times, while<br/>maintaining a reasonable accuracy. We will<br/>discuss two of such model types: (1) conceptual<br/>models and (2) machine learning models.