Typically, hail contamination of radar rainfall estimates is simplistically handled by a maximum rain intensity threshold, e.g., 103.8 mm/h in case of the NWS. However, this threshold can vary widely depending on local conditions. For example, Fulton et al. [1998] reported values ranging between 75 and 150 mm/h. Mendez et al. [2003] analyzed maximum gauge intensities at different durations and frequencies for summer thunderstorms over Walnut Gulch. They found that the maximum 5-min intensities (i.e., at the radar data time step) frequently exceeded 103.8 mm/h, and sometimes were near 250 mm/h. Using radar data, Morin et al. [2006] modeled rain cells as circular Gaussian elements with a maximum intensity and a decay factor: these can give theoretical maximum pixel-scale intensities greater than 103.8 mm/h. From the latter two studies, and because flash flood forecasting mainly deals with intense rainfall over short return periods, we raised the hail threshold to a reasonable 146.3 mm/h, which has been estimated as the mean of the maximum 5-min gauge intensities for a 10-year return period [Mendez et al., 2003]. Note that the actual WG11 peak storm intensities are rarely close to, and are mostly less than, this new value (Table 2). However, peak storm intensities much larger than 146.3 mm/h have been observed over the entire WGEW. Raising the hail threshold provided a better match between the cumulative basin-average radar and gauge rainfall series.