On the usefulness of weather radar data in real-time hydrological forecasting in Belgium

Describing rainfall-runoff relationships a linear stochastic black-box modelling technique is chosen, yielding what is known as a transfer function noise model. An important parameter in this representation of the real system is the dead time or concentration time of the system. This parameter directly influences the forecasting horizon. Efforts to extend this forecasting horizon are concentrated on the development of procedures to estimate future precipitation. Research can be directed along two main directions: statistical analysis (e.g. rainfall generators) and forecasting techniques based on meteorologic and weather radar information. Assuming that the error made in estimating precipitation intensities is acceptable, hydrological forecasting can benefit from this information. As a first attempt to incorporate radar information, one can try to develop some decision making procedure to choose between rainfall scenarios. A simulation study is used to appreciate the usefulness of incorporating weather radar data in the flood forecasting model of the river Meuse. -from Authors