No abstract

No abstract

SUMMARYThe internal circulation and the shape of water drops falling at terminal velocity in air of 20°C at sea level pressure, and nearly water saturated, were studied by means of a wind tunnel. Drops with an equivalent radius a, smaller than 140 pm had within the experimental error no detectable deformation from spherical shape.Drops of sizes 140 pm < a. < 500 p m were slightly deformed into an oblate spheroid. The deviation of these drops from spherical shape was found to be in fair agreement with that theoretically predicted by Imai (1950) and others. The deformation of drops of sizes 0.5 mm < uo < 4.5 mm was found to be linearly related to the drop size. Such a linear relationship is predicted by the semi-empirical calculations of Savic (1953).By means of a tracer technique it was established that water drops falling at terminal velocity in air have a well developed internal circulation. T h e flow pattern inside a drop was found to be consistent with the flow pattern of the air around the drop and that predicted theoretically by Hadamard (1911) and by Hamielec and Johnson (1962). The surface velocity at the equator of a drop was found to be about 1/100 of the drop's terminal velocity. The experimentally determined internal velocities were compared with those predicted theoretically by McDonald (1954) from boundary layer theory and by Hadamard (1911) based on Stokes flow.

The oscillations of moderate to large raindrops are investigated using a seven-story fall column with shape data obtained from multiple-strobe photographs. Measurements are made at a fall distance of 25 m for drops of D ϭ 2.5-, 2.9-, 3.6-, and 4.0-mm diameter, with additional measurements at intermediate distances to assess the role of aerodynamic feedback as the source of drop oscillations. Oscillations, initiated by the drop generator, are found to decay during the first few meters of fall and then increase to where the drops attained terminal speed near 10 m. Throughout the lower half of the fall column, the oscillation amplitudes are essentially constant. These apparently steady-state oscillations are attributed to resonance with vortex shedding. For D ϭ 2.5 and 3.6 mm, the mean axis ratio is near the theoretical equilibrium value, a result consistent with axisymmetric (oblate/prolate mode) oscillations at the fundamental frequency. For D ϭ 2.9 and 4.0 mm, however, the mean axis ratio is larger than the theoretical equilibrium value by 0.01 to 0.03, a characteristic of transverse mode oscillations. Comparison with previous axis ratio and vortex-shedding measurements suggests that the oscillation modes of raindrops are sensitive to initial conditions, but because of the prevalence of offcenter drop collisions, the predominant steady-state response in rain is expected to be transverse mode oscillations. A simple formula is obtained from laboratory and field measurements to account for the generally higher average axis ratio of raindrops having transverse mode oscillations. In the application to light to heavy rainfall, the ensemble mean axis ratios for raindrop sizes of D ϭ 1.5-4.0 mm are shifted above equilibrium values by 0.01-0.04, as a result of steady-state transverse mode oscillations maintained intrinsically by vortex shedding. Compared to the previous axis ratio formula based on wind tunnel measurements, the increased axis ratios for oscillating raindrops amount to a reduction of 0.1-0.4 dB in radar differential reflectivity Z DR , and an increase of about 0.5 mm for a reflectivity-weighted mean drop size of less than about 3 mm.

Previous calculations of the rate at which falling droplets in clouds collide with aerosols have led to the conclusion that except in thunderclouds any electrical charges on the aerosols or droplets have little effect on the collision rate. However, it had been assumed that the aerosols would have only a few elementary charges on them, whereas it is now known that at the tops of nonthunderstorm clouds the evaporating droplets may have several hundred elementary charges on them and that much of this charge remains on the residual aerosol for 5 min or so after the evaporation. Also, most previous calculations neglected image charge forces that provide strong attraction at close range even when droplet and aerosol have charges of the same sign and of comparable magnitude.The authors present numerical calculations showing that electrical effects dominate collision rates for charged evaporation aerosols. The calculations are for the size range of 0.1-to 1.0-m radius with the collision efficiency compared to that for phoretic and Brownian effects being greater by up to a factor of 30 greater for droplets from 18.6-to 106-m radius with relative humidity in the range 95%-100% and only 50 elementary charges on the aerosol. The results imply that electrical effects can be important for the scavenging of evaporation aerosol particles in the size range of the Greenfield gap.The authors call this process ''electroscavenging.'' Electroscavenging of charged particles, when the particles are mostly of the same sign, is a previously unrecognized droplet charging process. Electroscavenging also provides a pathway for contact ice nucleation when charged aerosol particles from evaporated charged droplets collide with supercooled droplets. Ice nucleation can occur because aerosol particles from the evaporation of cloud droplets have been found to be more effective as ice forming nuclei than other aerosol particles that have not been processed through droplets.