A Field Study of Pixel-Scale Variability of Raindrop Size Distribution in the Mid-Atlantic Region

The spatial variability of parameters of the raindrop size distribution and its derivatives is investigated through a field study where collocated Particle Size and Velocity (Parsivel 2) and two-dimensional video disdrometers were operated at six sites at Wallops Flight Facility, Virginia, from December 2013 to March 2014. The three-parameter exponential function was employed to determine the spatial variability across the study domain where the maximum separation distance was 2.3 km. The nugget parameter of the exponential function was set to 0.99 and the correlation distance d(0) and shape parameter s(0) were retrieved by minimizing the root-mean-square error, after fitting it to the correlations of physical parameters. Fits were very good for almost all 15 physical parameters. The retrieved d(0) and s(0) were about 4.5 km and 1.1, respectively, for rain rate (RR) when all 12 disdrometers were reporting rainfall with a rain-rate threshold of 0.1mmh 21 for 1-min averages. The d(0) decreased noticeably when one or more disdrometers were required to report rain. The d(0) was considerably different for a number of parameters (e.g., mass-weighted diameter) but was about the same for the other parameters (e.g., RR) when rainfall threshold was reset to 12 and 18 dBZ for Ka- and Ku-band reflectivity, respectively, following the expected Global Precipitation Measurement mission's spaceborne radar minimum detectable signals. The reduction of the database through elimination of a site did not alter d(0) as long as the fit was adequate. The correlations of 5-min rain accumulations were lower when disdrometer observations were simulated for a rain gauge at different bucket sizes.