General Principles

Flood forecasting models are an important component in many flood warning and emergency response systems. Models can assist by providing advance warning of the likely timing and magnitude of flooding, and in helping to understand the complexities of a flood event as it develops. Models outputs may also be used in decision support systems for flood event management and the operation of flow control structures. The techniques used for flood forecasting have many similarities to the methods used for simulation modelling of river and coastal processes. However, the design may be constrained by the availability of real time data, and computer systems on which to operate the model, although there is the advantage that model outputs can be updated to help to account for differences with observed values; a process which is often called data assimilation. Ensemble and probabilistic techniques are also increasingly being developed to provide information on model uncertainty to users of model outputs, and to allow a more risk-based approach to decision making. This chapter provides a general introduction to these issues and to the topic of flood forecasting model calibration and performance monitoring.