Modelling runoff in an urban catchment using Vflo™ and QPE

Real-time flood forecasting has benefited greatly by the application of physically based hydrological models and real-time rainfall data obtained from weather radar. An application of a distributed hydrological model (Vflo™) to an urban catchment using distributed rainfall data from Nexrad radar images illustrates some important issues. The benefits of using Nexrad data with pixels approximately 1 km² as opposed to using a single raingauge located at the edge of the catchment are demonstrated through model simulation. It was also possible to improve the model results by developing a new conceptual model of the 10·25 km² Harris Gully catchment (in Houston, Texas) and replacing an earlier model calibrated on the much larger Brays Bayou catchment. In the case of the Harris Gully catchment, flooding has been caused by several factors including heavy rainfall on the catchment exceeding the design capacity of the existing storm sewer system, and backwater effects caused by the larger Brays Bayou (which Harris Gully flows into) running at full capacity. Improving the existing flood forecasting system (FAS, the flood alert system) for Harris Gully is of high priority since the catchment contains the Texas Medical Center and the main campus of Rice University. It is planned to do this through the installation of a state-of-the-art NetRAD weather radar in the near future.