Many arid and semi-arid regions are particularly affected by flash floods, caused mainly by convective storm systems, and often resulting in significant damages to life and property. The short duration and spatially compact nature of these storms, along with the highly nonlinear nature of arid and semi-arid watershed repsonse, makes forecasting and predicting subsequent flash-floods extremely difficult. In an attempt to improve our flash-flood prediction capability, a semi-arid specific model based on the well-established event-based rainfall-runoff model KINEROS2 has been developed which utilizes National Weather Service Doppler radar Digital Hybrid Reflectivity product. KINEROS2 is a spatially distributed kinematic wave model which represents the basin as a cascade of overland flow and channel model elements. The dynamic infiltration algorithm in KINEROS2, used to treat both hillslope and channel transmission losses, is particularly well suited for simulation of semi-arid hydrologic processes. KINEROS2 was restructured and coupled with optimization tools for parameter estimation. This presentation will briefly describe the model and data processing developments. More importantly, it will assess the lessons learned in implementing and applying this flash-flood forecasting system in the Tucson National Weather Service office during the 2006 monsoon season.