Flood modelling and forecasting are core parts of flood risk management. The last decade has seen rapid advances in the development of efficient and flexible flood 
simulation frameworks for applications ranging from river-reach up to the global scales.

Flood modelling uses two-dimensional hydrodynamic simulators with as much complexity as needed, depending on the length of the spatial (kilometre to submeter) and temporal 
(day to microsecond) scales and on the level of detail in the outputs. Complex simulators can be useful to capture the details of the hydrodynamic processes involved in 
application at river-reach and neighbourhood scales. However they are computationally expensive and are unlikely to outperform simplified simulators for large-scale 
fluvial/pluvial applications.

Flood forecasting is affected by multiple sources of uncertainty that need to be efficiently quantified and reduced. Uncertainties stemming from observational and forcing 
data, model parameters, the initial state of the system, and flood model structure can affect the overall accuracy. Still, it is challenging to have a general and 
efficient framework to identify the relevant uncertainty sources and how their joint effects impact the output flood-related probability distributions.