The raindrop size distribution (DSD) plays an important role in determining the radar measurements. From the radar perspective, the accurate parameterization of the tail of the DSD is important because the radar parameters are strongly influenced by the concentration of the large drops or on the tail of the DSD (Bringi and Chandrasekar, 2001). In the issue of rainfall estimation by radar through the necessary relationship between radar measurement and rain rate, the main limitation is attributed to the variability of the radar rainfall relationship. And it is well known that the sensitivity to natural variation in DSD is a substantial source of the relationship's variability (Maki et al., 2005). The raindrop number and size at ground are being the resultant of various microphysical processes, a suitable approach would need to analyze their variability in relation to that of rainfall estimation relationship.
In this study, an effect of DSD variability on radar measurements according to the maximum diameter was investigated for the purpose of determining the variability of rainfall estimation relations. The shape of averaged drop size distribution was produced as a function of the maximum diameter (Dmax): The shape of small Dmax had upward vortex shape but it was changed to linear shape as Dmax increased. The Dmax played a role as a fingerprint to present the variability of DSD: shape m, slope Λ, and intercept parameter N0 generally decreased with increasing maximum diameter and median volume diameter D0 and mass weighted mean diameter Dm values are linearly increased with Dmax. The characteristics of DSD variation according to the Dmax were affected to the rain rate estimator. On the Z-R relation (Z=ARb), the coefficient A increased as Dmax increased whereas the exponent b decreased. On the R-KDP relation (R=aKDPb), coefficient a decreased and exponent β increased as Dmax increased.
Based on the analysis of implication of DSD variability in regards to rainfall estimation relationship variability, rain rate estimation algorithm was developed considering DSD variability to make optimized empirical relationships of coefficient and exponent on the rain rate estimator with Dmax. The new rain rate estimation algorithm reduced RMSE (root mean square error) and normalized error of the results as compared with conventional rain rate estimator.