WMOD00420 - Laserfiche WebLink

Home Browse Search

558 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 5 where F" Fs are the target and free-stream particle flux, A" As the target and free-stream sample area, and ~, Vs the target and free-stream particle speed. The ideal sample area for infinitesimal particles at the mouth of the aspirator, As. is defined by the four converging streamlines which terminate on the comers of A" the target, or 2D-C sample area in the throat of the aspi- rator. This is the sample area assumed by the constant air flux, A,Va = 0.66 L S-I, used in the concentration measurements presented here, where Va is the airspeed at the target. As particle momentum increases, F, should decrease since fewer particles in As will be entrained into A, by the convergent airflow. Thus the result, assuming V, = Va, should be a decreasing C" and hence Cm, as a function of size with Cm - 1.0 for the smallest sizes. This prediction is not corroborated by the measure- ments presented here nor by the calculations of Nor- ment (1985), and the reason can only partly be ascribed to ignoring the differences of V, and Va. For the mea- surements here on ice particles, ~ was thought to be 8 m S-I = 0.6Va (Holroyd 1986). This suggests Cm should decrease from -1.7, whereas values ofCm > 3.0 were observed. From Norment's calculations of the ra- tio F,:Fs, Cm was evaluated assuming ~ = Va. The results indicate that Cm varies from 1.0 to 3.1 for water droplets 20 to 300 JLm, and Cm < 1.0 for droplets> 300 JLm; whereas if the correct ~ is used, Cm varies from 1.0 to 3.9 for 20 to 400 JLm droplets, and Cm < 1.0 for droplets> 400 JLm. It appears that particle trajectories cross streamlines to become more concentrated at the target. Norment explains this as a result of the bending of free-stream particle trajectories as they enter the as- pirator, since the aspirator axis is not parallel with the free stream trajectories. This result appears to be borne out by the measurements presented here. Indeed, Nor- ment's calculations of Cm for a horn axis of 600 are quite similar to the results presented in Table 3. For ice particles entrained into the converging air- flow, the quantity V,/Va will be primarily a function of particle mass and independent of winds; whereas F,/ Fs will depend on the total momentum of the par- ticle, since particles with higher momentum will be less likely to follow the convergent airflow. Thus, as crosswinds increase, F,/Fs, hence Cm, should decrease. This result is borne out by Fig. 11, displaying Cm as a function of size from comparisons with the truckborne 2D-C; although, results from comparison with the air- borne 2D-C are not consistent with this prediction. Another point which should be mentioned is that the fixed probe with a vertical aspirator functions sim- ilarly to an unshielded precipitation gauge. As such, the efficiency with which snowfall is sampled will be a function of windspeed. The higher the winds, the lower the efficiency. The sampling efficiency of the aspirator may be slightly higher than a precipitation gauge be- cause of light suction at the top of the aspirator, but this would be important only for smaller particles. Two of the three empirical methods used here for ground truth do not have a wind dependent sampling effi- ciency, while the collection efficiency of the third, oil slides, is primarily determined by the 8.8 m S-I speed of the slides at the end of the rod. Differences caused by wind would primarily cancel themselves, since col- lection efficiency would increase as the slide moved against the wind and decrease as the slide moved with the wind. Thus, these correction factors account for both the effects of aspiration and the wind dependent collection efficiency of the unshielded aspirator horn. 5. Summary and conclusions Application of particle imaging technology to snow- fall measurements is difficult because of the require- ment to know the speed of particles passing through the imaging laser beam. Attempts to solve this problem by artificially accelerating ice particles, with suction fans at the base of a parabolic horn which fits over the probe, have met with limited success since the errors caused by aspiration have been unknown. Motivation for calibration of such an instrument is provided by the fact that it can provide high resolution snow crystal concentration and size distribution measurements in a form that can be easily managed with computer pro- cessing. Three empirical methods to compare with snowfall measurements made using an aspirated 2D- C have been presented here. The aspirated 2D-C was fixed vertically on the ground and simultaneous measurements were made with a 2D-C on an aircraft making low passes overhead, or an oil-coated slide spun on the end of a rod, or a 2D-C carried on a truck. Comparisons ofIPC obtained with the aspirated probe and the other three techniques are reasonably well correlated and in rough agreement. The IPC measured by the aspirated 2D-C are higher by factors of 3.2 compared to the airborne 2D-C, and 2.4 compared to the oil slides and truckborne 2D-C. For the airborne and truckborne 2D-C measure- ments, snowfall size distributions were also compared with the aspirated measurements. For each comparison Cm, the ratio (aspirated:unaspirated) ofIPC by size bin was calculated. For the aircraft data, four comparisons were made, while for the truck-mounted unit 82 com- parisons were available. These comparisons produced a reasonably consistent picture, indicating that Cm de- creases as particle size increases, becoming < 1.0 for large particles. This tendency is in qualitative agree- ment with Norment's (1985) calculations. Comparison with the truckborne 2D-C indicated Cm < 1.0 for par- ticles > 1 mm, while for the airborne 2D-C Cm < 1.0 for particles> 0.7 mm. Further stratification by wind- speed of comparisons with the truckborne 2D-C shows Cm decreasing for all particle sizes as winds increase. Comparisons with the airborne 2D-C, however, do not agree with this tendency. In near calm conditions, Cm was still found to vary as a function of particle size - :