Raindrop Size Distributions and Z–R Relationships in Southeast China: Hilly Inland vs. Coastal Plain Sites

Knowledge of raindrop size distribution (RSD) is essential for understanding microphysical processes occurring within cloud and precipitation systems, as well as for enhancing the capabilities of numerical models and radar-based quantitative precipitation estimation (QPE). However, observation and study of RSD, especially its temporal and spatial variability, remain quite limited in specific regions. One such region is Southeast China. In this paper, four years of disdrometer data from a south coastal plain site (CPS) and a north hilly inland site (HIS) in the Fujian Province of Southeast China are analyzed and compared to elucidate the characteristics and discrepancies of RSD between these two distinct climatological sites. On this basis, empirical relations between the parameters of Gamma distribution and between radar reflectivity factor (Z) and rain rate (R) are proposed. The results are summarized as follows. (1) In the cases of light to moderate rains, HIS exhibits a higher (lower) concentration of small-size (midsize and large) raindrops with diameters of D < 1 mm (1 $⩽$ D < 3 and D $⩾$ 3 mm), compared to CPS. Conversely, as the rain intensity increases, the raindrop concentrations across all size categories at CPS gradually exceed those at HIS. (2) RSDs at both sites broaden and exhibit elevated concentrations across most diameter categories as the rain rate increases. (3) For rainfalls with rain rates below 5 mm h⁻¹, collision and coalescence dominate, resulting in unimodal rain spectra at both sites; whereas for stronger rainfalls, breakup intensifies, leading to the development of bimodal rain spectra. (4) HIS experiences more stratiform rains but fewer, weaker convective rains than CPS. Stratiform RSD at HIS possesses more small and large raindrops but fewer midsize raindrops compared to CPS, whereas convective RSD at CPS possesses higher concentrations across all diameter categories. (5) Accordingly, specific Z–R relations at these two sites are proposed and validated for two real cases, demonstrating that the accuracy of radar QPE is effectively improved based on the proposed Z–R relations.