Sydney, Australia, has a stormwater system that is completely separate from the sewerage system. The stormwater systems are mainly managed by the 43 local councils and most of these use OSD as a means of reducing site discharge of stormwater, thereby relieving catchment flooding. Sheas Creek is an urbanized catchment that lies within the South Sydney City Council Local Government Area. Since 1984 an OSD policy has been implemented in the Sheas Creek catchment. The associated design code, like most others in Sydney, is based on discharge from individual properties without consideration of runoff from other parts of the catchment. This design code is therefore not based on a total catchment approach. The objective of this study is to assess the effectiveness of OSD in a catchment-wide manner. More specifically, to see how existing and planned OSD affects flooding in the Sheas Creek catchment, and to test the effectiveness of the current rate of permissible site discharge (PSD). The study also looks at the effectiveness of OSD when different spatial distributions are used across a catchment. The methodology involved using a computer rainfall-runoff model, OSDSAX, which is based on the ILLUDAS model. This was configured for the Sheas Creek catchment and was calibrated to match flood level data recorded during four large storm events. The model was then adopted to investigate the impact of OSD. Various scenarios were tested to satisfy the objectives of the study. Some of the results challenge conventional wisdom and these arise from catchment-specific characteristics. The analysis shows that the perceived beneficial effects of OSD are in some cases unrealistic and that catchment modelling is required to test and quantify the actual impacts for specific catchments.