A new understanding of raindrop shape

Abstract

In this paper raindrop shapes from laboratory, field and model investigations are examined in order to distill a consistent picture of raindrop axis ratios as a function of size. The first realistic measurements of raindrop shape were made in the UCLA vertical wind tunnel that was built under the supervision of Hans R. Pruppacher at UCLA in 1968. Subsequent research has investigated raindrop shape in more detail by laboratory studies, field measurements and modeling. Measurements of rain at night in backscattered light have revealed that raindrops larger than 1 mm diameter oscillate in conditions where winds and collisions are too weak to excite oscillations. One apparent source of these oscillations is a resonance with eddy shedding. Subsequent laboratory studies have demonstrated that such oscillations increase the average axis ratio of small raindrops. Details of past theoretical research show that raindrop axis ratios decrease with size in the manner of a sessile drop because of its increasing weight, but also because of an increasing differential in the fore-aft aerodynamic pressure. Modern analytical and numerical models predicted similar axis ratios for water drops falling at terminal velocity and these results are generally consistent with laboratory and field measurements. Although the axis ratios first measured in the UCLA wind tunnel and the more recent laboratory measurements have led to a broader understanding of raindrop axis ratios, new information about raindrop shape is available from recent field measurements using improved techniques. Raindrop axis ratios, orientation and canting angles using 2-D video disdrometer (2DVD) are also reported. The results include measurements from an artificial rain experiment conducted under calm wind conditions where the drops (over 115,000) were allowed to fall a distance of 80 m. Axis ratios and drop shapes show good agreement with the Mainz wind-tunnel data and images, as well as canting angle distributions symmetric about zero degrees with a standard deviations of 7–8° for all drops > 2 mm. Measurements of 2DVD raindrop shape has been collected from Okinawa, Sumatra, Toronto, Huntsville and Brisbane. An unusually wide axis ratio distribution was observed in the D = 1.6–1.8 mm region, which may be evidence for drop oscillations caused by eddy-shedding. Some subtle differences can also be seen for larger drops (e.g. D = 4 mm), the vertical chord being slightly larger than the horizontal chord leading to slightly larger axis ratios compared with the 80 m fall data. Additionally, an example using individual drop information from 2DVD measurements is considered, which enhances agreement with C-band polarimetric radar measurements. In this particular case, a considerable shift in the axis ratio distribution for 3.5 mm drops is observed during the initial strong convective phase of the event. Further ongoing work will combine the 2DVD measurements with polarimetric radar observations in a synergistic manner in order to identify the meteorological or microphysical conditions under which significant changes in drop shapes may occur in natural rain. Improved polarimetric radar estimates of canting angles are reported from an S-band system with high cross-polar performance antenna. The histogram of the standard deviation derived from the radar data obtained in light stratiform rain event with embedded convection shows a canting angle mode of about 7°, with a significant positive skewness. An increase in stability with drop diameter is also inferred from the radar measurements, which is in agreement with direct estimates from the 2DVD. Ongoing experiments in the Mainz wind tunnel are evaluating how the raindrop shapes are affected by collisions with drizzle drops within rain shafts.