The retrieval of raindrop properties based on a Constrained Gamma model using dual-polarization radar and 2-Dimenstional Video Disdrometer in Korea

The raindrop size distribution (DSD) is a very useful for understanding rain microphysics and improving microphysics parameterization in numerical weather prediction (NWP) models (Steiner et al., 2004). The DSDs are associated with the precipitation type, weather condition and storm developments, and can be characterized through disdrometer and dual-polarization radar measurements. The observed rain DSDs from the disdrometer can be represented by various DSD models such as M-P, exponential, gamma, and Constrained-Gamma (C-G) DSD model. Among them, C-G distribution model based on empirical relation between shape and slope can be used as a constraint that allows DSDs to the retrieved from reflectivity (Z) and differential reflectivity (ZDR) of dual-polarization radar (Zhang et al., 2001). The shape-slope relation is derived by disdrometer data. It reduces the gamma DSD model to a two parameter DSD model. However, DSDs are variable in specific regions and the shape-slope relation may vary in different climate regions (Cao et al., 2008). In order to achieve more accuracy rain DSDs, shape-slope relation assumption reflecting rainfall characteristic of Korea is needed. For retrieval of raindrop properties by reflecting Korea's rainfall characteristics, we derived new shape-slope relation using 2-Dimensional Video Disdrometer (2DVD) data from April 2014 to March 2015 (rainfall cases) in Jincheon, Korea. The 2DVD measures the DSD of precipitation including drop size, fall velocity and oblates of individual raindrops. The 2DVD data is useful to investigate the characteristics of rainfall. However, the number of particles outliers measured by break up of drops, wind turbulence and so on (Raupach and Berne, 2015). Therefore, we apply to the velocity-based filter for remove drop outliers, and compared the rain rate calculated from the 2DVD observation to collocated rain gauges measurement. And the 2DVD data of less than 5 channels was removed for the continuity of the data. Gamma distribution is consists of three governing parameter such as number concentration (No), shape and slope parameters. These parameters are calculated from the second, fourth, and sixth moments (Zhang et al., 2001). To obtained empirical shape-slope relation, total 42 rainfall cases (stratiform (29), convective (6), mixed (7)) with 27,593 minute DSD samples are used. We apply to condition for drop counts greater than 1,000 min-1 and rain rates greater than 5 mm hr-1 (Brandes et al., 2003). The derived shape-slope relation by 2DVD data with existing relation by Zhang et al. (2001) and Cao et al. (2008) are compared in Figure 1. The solid line is new empirical shape-slope relation dotted line represents the Florida shape-slope relation from Zhang et al. (2001) and dashed line represent the Oklahoma relation from Cao et al. (2008). As the slope increased, shape-slope relation differed. It can be seen as the regional difference in precipitation characteristics. Its impact on the retrieved rain DSD may not be neglected, implying Korean rainfall characteristics should be reflected in the C-G DSD model. To verify the derived shape-slope relation, we examine the mass-weighted mean diameter (Dm) and the standard-deviation of Dm (óm). The results show that, the Dm-óm relation derived from the 2DVD observation corresponds well with from the new shape-slope relation. As a result of comparing the retrieved Dm and óm using new shape-slope relation and observed Dm and óm, the correlation coefficient (Corr), RMSE and BIAS of Dm parameter is about 0.87, 0.18 mm and -0.033 mm, respectively. The óm showed the similar statistics, with Corr=0.87, RMSE=0.09 mm and BIAS=0.017 mm. These results show that, the new shape-slope relation is valid for the rain properties retrieval. We will retrieve raindrop properties using new shape-slope relation and reflectivity and differential reflectivity of dual-polarization radar in Yongin, Korea. The retrieved DSDs are verified by comparison with 2DVD data. Also, we will show how the shape-slope relation effects can caused the difference in the retrieved rain DSD, when we use new shape-slope relation of rainfall derived from the Korean data. Acknowledgements: This research is supported by "Development and application of Cross governmental dual-pol. radar harmonization (WRC-2013-A-1)" project of the Weather Radar Center, korea Meteorological Administration.