The disdrometer, raingauge network and dual-Doppler radar data were collected during Typhoon Nari landfall. Three different methods are used to calculate the radar rainfall estimate. First, from radar data the stratiform and convective regions from both the reflectivity field and vertical motion field are decided. Each separation map is compared to the rain gauge data. In both the stratiform and convective regions, the Z-R tables through probability matching method for the plain area were derived. A second set of rain gauge data is used to verify the area rainfall.

The second approach is based on the disdrometer observation. The Gamma drop size distributions are derived from disdrometer observation at six minutes interval. The Z-R relations through Gamma distribution are calculated. During the heaviest rainfall period of Nari, the disdrometer observation indicated the maximum drop size did not exceed 4mm. The preliminary results showed that the N0 increased as the rainfall rate increased. The Z-R relation with the average of the coefficients derived from drop size distributions is Z = 150 R1.4 . The rainfall estimate from the disdrometer derived Z-R relations will be compared to the probability matching method results.

The Central Weather Bureau in Taiwan has been practiced a sequence of quality control of the reflectivity data. An optimum rainfall algorithm using both the radar and gauge data through objective analysis has been carried out for operation. The rainfall estimate will also be compared to the former two methods. Due to the complex terrain of northern Taiwan, the locations of the underestimate of the reflectivity by the blockage and partial beam filling have been identified. The vertical profiles of reflectivity in both stratiform and convective regions will be studied for the purpose of better extrapolation of reflectivity at lowest level. The derived reflectivity from drop size distribution will also be compared to the radar observation.

.