Flash floods caused by convective rain storms are highly sensitive to the space-time characteristics of rain cells. In this study we exploit the high space-time resolution of the radar data to study the characteristics of the rain cells and their impact on flash flood magnitudes. A rain cell model is applied to the radar data of an actual storm and the rain fields represented by the model further serve as input into a hydrological model. Global sensitivity analysis is applied to identify the most important factors affecting the flash flood peak discharge. As a case study we tested an extreme storm event over a semi-arid catchment in southern Israel. The rain cell model was found to simulate the rain storm adequately. We found that relatively small changes in the rain cell's location, speed and direction could cause a three-fold increase in flash flood peak discharge at the catchment outlet.