The high storm depth biases in the radar estimates (see section 2.1) made the WG11 gauge data with a 0.01″ accuracy [Keefer et al., 2008] more reliable. Following the “bulk adjustment” optimization approach [Krajewski and Smith, 2002], the event-specific radar rain estimates were improved by scaling to match the basin-average storm totals against those from the gauges, and by adjusting the radar-gauge time lag. These event-specific lags were calculated as the values at which maximum correlation of the WG11 average storm depth series occurred (Table 2). The gauge and improved radar estimates were merged through simple averaging, since there was no reason to prefer one spatial distribution over the other. Figure 2 shows the pattern similarity in some individual and merged cumulative hyetographs.