Abstract
This paper describes the development and operation of the new flood forecasting system for the Demer basin in Belgium, which has suffered severe flooding problems for many years. The system takes telemetry data from a large number of hydrological, meteorological, and hydraulic observation sites across the basin, along with radar rainfall forecasts. These data are fed through a network of complex hydrological and hydrodynamic models. Scheduled runs of the system take place at frequencies determined by the level of alert; runs are performed on a cluster of server computers, then results are available on client computers in the control room and via remote access. This gives operational basin managers fast, accurate, real‐time flood forecasting and flood mapping based on high‐resolution digital ground model information, enabling identification of those streets and areas affected by flooding. The basin managers have several options for diverting flood waters in order to avoid or mitigate the effects of flood events in key areas. The forecast results underpin a decision‐support system, giving basin managers the information they need to make informed judgements for disseminating flood warnings and altering their management of the river controls.