A multi-scale, physically-based flash flood prediction system has been developed for the Front Range region of Colorado which capitalizes on recent numerical weather prediction, nowcasting and hydrological modeling developments as well as high-performance computing capabilities. The system, operating in real-time during the summer of 2008, provided coupled weather and hydrological model forecasts of rainfall and streamflow, respectively. Convection permitting (1 km) Weather Research and Forecasting (WRF) numerical weather prediction (NWP) model forecasts were fed into the Noah-distributed hydrological modeling system to create hydrometeorological forecast guidance products, or regional assessments of potential severe weather and flooding threats. Additionally, 0-1 hour radar-based precipitation nowcasts were also generated operationally and were fed into the hydrological model. In this work we summarize various performance aspects of the coupled forecasting system. Emphasis is placed on the evaluation of the precipitation forecasts and nowcasts using a mosaic of local NEXRAD radars as well as opportunistic scans of the Colorado State University, dual-polarized CSU-CHILL radar. Several features of significant rainfall events are evaluated including storm intensity, duration, track as well as more regionalized metrics commensurate with general ‘guidance' forecasts. Selected hydrological forecasts are also evaluated in terms of their skill in forecasting the observed flood responses from heavy rainfall events. Each forecast product will is also evaluated in the context of its respective forecast lead-time and recommendations are provided as to future improvements of the system as well as future operational development.