The MIKE FLOOD modelling package (2012 release) was used to develop and calibrate a hydraulic flood model for the Punrak catchment (located approximately 56 km south of Perth, Western Australia). The model was developed in a coupling environment where 1-D MIKE 11 channel flows interact with the 2-D MIKE 21 overland flows. The interactions take place via a series of well-defined lateral linked cells and standard linked cells. The lateral linked cells run along both banks of the channels/drains. Standard linked cells are located next to the upstream ends of the channels/drains and both ends of a culvert. Rainfall is applied on each computational cell and the runoff is routed within the hydraulic model. In this study, the developed model was calibrated to the 1996 winter event at two gauged locations of the model domain, one at the centre (Hopelands Rd gauging station) and the other at the outlet (Yangedi Swamp gauging station) of the catchment. The predicted results in terms of peak discharge, peak flood level, time to peak and volumes were compared with the observed data. There was fair agreement between predicted and observed results. The peak flood level and total storage volume at the outlet of the catchment were underpredicted by 0.5% and 1.2%, respectively, whereas, the peak discharge was overpredicted by 2.4%. The model results were obtained based on a 10 m grid resolution. A finer grid resolution would have improved the modelling results. It was not considered in this study as this would increase the computational cost significantly. This would also affect the project budget and time-frame. This paper explores the application of the direct rainfall technique, a comprehensive model setup of the MIKE FLOOD model for the study area and detailed comparisons of the calibration results.
