SEARCH

SEARCH BY CITATION

Keywords:

  • ice crystals;
  • mixed-phase cloud;
  • ice particle growth

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References

[1] The characteristics of the capture and freezing of supercooled water droplets by falling ice crystals were quantitatively investigated in a vertical supercooled cloud chamber of 10 m length at temperatures from −5 to −25°C. We present evidence that ice crystals can start the riming process at smaller sizes than previously reported. The results obtained show that the minimum dimension of ice crystals involved in riming is around 60 μm in diameter for hexagonal plates and 30 μm width and 60 μm length for columnar ice crystals. No substantial difference has been observed between the size distribution of the droplets accreted on the crystals and the cloud droplet size distribution, indicating that this process should not produce significant changes in the cloud droplet spectrum. The time elapsed between the ice crystal nucleation and the collection of crystals with droplet accretion was typically around 60 seconds.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References

[2] Ice crystals in clouds grow by three different mechanisms: vapor deposition, collection of supercooled droplets (riming) and aggregation. In clouds consisting of a mixture of ice crystals and supercooled droplets, the observations indicate that after formation on ice nuclei, individual ice crystals must grow by vapor deposition to at least a critical minimum size before they can begin to collect cloud droplets. The time elapsed between ice crystal nucleation and the start of droplet accretion is dependent on crystal size and shape, and on the presence and size distribution of supercooled droplets in the cloud [Pruppacher and Klett, 1997; Young, 1993].

[3] The capture and subsequent freezing of supercooled cloud droplets by falling snow crystals is a very efficient mechanism of ice crystal growth which leads to the production of precipitation in the form of rainfall or snowfall. This process can modify the spatial extent, spatial distribution, and lifetimes of clouds [Pruppacher and Klett, 1997; Young, 1993], and is also an important process in the wet removal of pollutants from the atmosphere [Finlayson-Pitts and Pitts, 2000] and in the cloud electrification process [MacGorman and Rust, 1998; Saunders et al., 2001; Saunders, 2008].

[4] Field measurements show that there exists a cutoff size of ice crystals and cloud droplet diameters below which riming does not occur. Ono [1969] collected and replicated individual ice crystals in natural clouds. No riming was found on columnar ice crystals with minor axes less than 50 μm whereas almost all crystals of minor axis larger than 90 μm were rimed. On the other hand, it was observed that riming is rare on plates <300 μm in diameter but is usual on crystals >400 μm. Harimaya [1975] analyzed snow crystals collected on the ground and found that the critical dimensions are 90 μm minor axes diameter for columnar snow crystals, 30–40 μm minor axes diameter for needles, 300 μm diameter for hexagonal plates and 900 μm diameter for dendritic ice crystals. He did not find cloud droplets with diameters less than 10 μm collected by snow crystals. Reinking [1979] also studied the onset of riming of snow crystals based on snow crystal data collected during winter storms. This study reports that the minimum lengths or diameters of crystals before they begin to rime are within the range 115–320 μm for plates, while columnar crystals must grow to a minimum of 30–36 μm (width).

[5] Kajikawa [1974] studied model and natural snow crystals in free fall, measuring their collection efficiency for cloud droplets. For the model systems, he used circular disks, hexagonal plates and broad branched shapes as models of crystals. The sizes of the models were between 1 to 7 mm; therefore his study cannot give information about the cutoff size of small ice crystals for riming. For natural hexagonal plate snow crystals, he also found that riming is rare on plates smaller than 300 μm diameter.

[6] Theoretical calculations of collision efficiencies of thin ice plates colliding with supercooled cloud droplets have been performed by Pitter and Pruppacher [1974] and Pitter [1977]. They assumed that the flow fields past hexagonal plates can be approximated by the fields produced by thin oblate spheroids, and in conjunction with a superposition method they calculated the trajectories of the water drops relative to the ice crystals. They found that minimum ice plate diameter for a non-zero collection efficiency is around 300 μm for droplets of any size. Similar studies of theoretical collision efficiencies for simple columnar ice crystals were calculated by Schlamp et al. [1975]. The flow field past an ice column was approximated by the field of an infinitely long cylinder. These authors found that the collision efficiency decreases to zero for cylindrical ice crystals if their minor axis diameter is between 47 and 65 μm.

[7] More recently, Wang and Ji [2000] determined the collision efficiencies for hexagonal plates, broad branched and columnar ice crystals by using the 3-D flow fields of the falling ice crystals determined by solving numerically the Navier-Stokes equations and solving the equation of motion for a cloud droplet under the influence of the flow field. Using this method, they determined that the riming cutoff size is 70 μm minor axis diameter for columnar ice crystals, 220 μm diameter for hexagonal plates, and 400 μm diameter for broad-branched crystals.

[8] This work is aimed to obtain further knowledge on the growth of ice crystals by riming in cloud, for this reason the growth process is reproduced in the laboratory under controlled conditions and closely examined. In this paper we present a study of the minimum dimensions of falling ice crystals necessary for the initial stages of capture and subsequent freezing of supercooled cloud droplets. The sizes of the collected droplets are analyzed and the basic crystal growth habits are investigated. The study is based on ice crystal data collected in a laboratory cloud. As far as we know this is the first laboratory work studying the initial stage of the riming processes.

2. Mixed-Phase Cloud Chamber

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References

[9] The present experiments were carried out at the University of Manchester, England, where the altitude is 42 m asl. The experiments were performed in a supercooled cloud chamber, which consisted of a vertical tube of about 10 m length and 1m inner diameter, this tube is placed inside three vertically stacked cold chambers with temperature control down to −50°C.

[10] The cloud of supercooled water droplets was generated by vapor condensation of water provided continuously by a boiler connected to a controlled power output in order to obtain the required liquid water content in the chamber. The air convection produced by the introduction of warm vapor into the base of the cloud chamber helped to mix and homogenize the temperature and liquid water content inside the cloud chamber. The temperature was continuously monitored by thermocouples placed close to the base, top and middle of the tube and the fluctuation (both, spatial and temporal) in air temperature during a run was typically around ±0.8°C on average. The liquid water content was estimated by weighing the deposit of rime ice formed on a rod of 3 mm diameter mounted in a tube with airflow of 10 m s−1. Liquid water content samples were taken from the bottom and middle of the cloud chamber in order to check the homogeneity of the water droplet cloud.

[11] The ice crystals were nucleated 1m from the top of the cloud chamber by cooling a local volume of the droplet cloud with a rapid expansion of air from a compressed air source.

[12] The following steps were followed to run an experiment:

[13] 1. The cold chambers were settled at the desired temperature.

[14] 2. The boiler introduced vapor into the base of the chamber for about 20 minutes. The droplets filled the cloud chamber, reached thermal equilibrium with the environment and the cloud had time to come to a steady water content.

[15] 3. The ice crystals were nucleated and the boiler continued introducing vapor up to the end of the run.

[16] 4. After nucleation, the ice crystals grew at the expense of the droplets in the chamber and descended under the influence of gravity and convective current, falling through the supercooled water droplet cloud.

[17] 5. The ice crystals were collected by exposing a glass-slide coated with a 3% formvar solution [Schaefer, 1956].

[18] The wall of the cloud chamber became frosted during the experiments but the number of ice crystals detached from the wall was negligible in comparison to the number of ice crystals produced by nucleation. Besides, frost fragments are clearly different in shape to the crystals.

3. Measurements and Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References

[19] The crystals from the replicas were analyzed by a combination of direct observations with a microscope and photographs taken with a camera attachment.

[20] Here we shall restrict our analysis and discussion to the collision-coalescence process between small, pristine, ice crystals and supercooled cloud droplets, which is applicable to the initial stages of riming process.

[21] Several ice crystal samples were taken during each experiment. In order to examine the minimum dimension of ice crystals involved in riming the size distribution of rimed crystals was compared with that of unrimed crystals.

[22] Figure 1 shows histograms of the major axis dimensions for collecting and non-collecting hexagonal plates at different stages of the run. Figures 1a, 1b, and 1c display the histograms obtained at 63, 99 and 200 seconds after nucleation, respectively. This experiment was performed at −22°C, the liquid water content before ice crystal nucleation was 1.7 g m−3 and the three samples were taken in the cloud chamber around 4 m below the nucleation region. The size distributions of the droplets accreted on the hexagonal plates are shown in Figure 1d together with the cloud droplet spectrum of the mixed-phase cloud, which was measured during some particular runs by passing very rapidly one slide coated with formvar across the mixed-phase cloud. Some images of hexagonal plates with droplet collection are shown in Figure 2.

image

Figure 1. (a–c) Size distributions of collecting and non-collecting hexagonal plates at different stages of the run. (d) Cloud droplet spectrum and the size distribution of the droplets accreted on the plates.

Download figure to PowerPoint

image

Figure 2. Images of rimed hexagonal plates collected at −22°C.

Download figure to PowerPoint

[23] It is interesting to note that hexagonal plates as small as 60 μm diameter are able to initiate the droplet collection process. The size distributions of the ice crystals change through the run; however, at a given time, the histograms show that the average diameter and size spread of rimed and unrimed hexagonal plates are similar, indicating that the collecting plates do not have a preferential size for collecting droplets. The percentage of rimed plates observed on the slides was between 20 and 40%.

[24] As can be seen, no marked difference is observed between the size distribution of the droplets accreted on the plates and the cloud droplet size distribution. No correction was made for spreading of droplets on the crystal surface. In fact, the diameters of the droplets were taken as measured on the photographs. In general, the observation of the degree of deformation is quite difficult, leading to considerable uncertainties. The minimum diameter of droplets collected by hexagonal plates is between 5 and 10 μm as shown in Figure 1d.

[25] The relationship between the length along the major axis (c-axis) and that along the minor axis (length along a-axis or width) of rimed and unrimed columnar ice crystals collected on the slides is shown in Figure 3. Figures 3a and 3b display the column sizes at 66 and 97 seconds after nucleation, respectively. The size distributions of the droplets accreted on the columns are shown in Figure 3c together with the cloud droplet spectrum of the mixed-phase cloud. Some images of columns with collected droplets are shown in Figure 4a. This experiment was performed at −6°C, the liquid water content before ice crystal nucleation was 2.1g m−3; both samples were taken in the cloud chamber around 4m from the nucleation region.

image

Figure 3. (a and b) Relationship between the sizes (length and width) of collecting and non-collecting columnar ice crystals. (c) Cloud droplet spectrum and the size distribution of the droplets accreted on the columns.

Download figure to PowerPoint

image

Figure 4. (a) Images of rimed columnar ice crystals collected at −6°C. (b) Images of rimed dendrites and sector plate crystals collected at −15°C.

Download figure to PowerPoint

[26] The ice columns collected on the slides range from 20–50 μm minor axis diameter and 40–140 μm major axis length. The results show riming on columnar ice crystals with width and length as small as 30 μm and 60 μm, respectively. The size distribution of cloud droplets collected on columnar ice crystals is very similar to that of the droplet cloud. The minimum diameter of droplets collected by columns is between 5 and 10 μm as shown in Figure 3c. The percentage of rimed columns observed on the slides was between 5 and 20%.

[27] For temperatures around −15°C the ice crystals collected have sector plate and dendritic habit types. They grow rapidly by vapor deposition because the growth rate has a peak around this temperature. Figure 4b displays images of rimed dendrites and sector plates collected on the slide 125 seconds after nucleation at −15°C. The average size of these crystals is around 250 μm diameter and the percentage of rimed crystals observed on the slides is around 90%. These samples were taken in the base of the cloud chamber around 9m from the nucleation region.

4. Discussion and Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References

[28] A number of microphysical and dynamical processes interact in the current experiments. A homogeneous concentration of supercooled cloud droplets fills the cloud chamber before the ice crystal nucleation. After nucleation, a mixed-phase cloud is formed which consists of a mixture of liquid droplets, water vapor and ice particles coexisting at the cloud chamber temperature. The ice crystals grow by vapor deposition at the expense of cloud droplets due to the difference of water vapor saturation over ice and water, so that this mixed-phase cloud evolves over time. This evolution can be seen in the different size distributions of the ice crystals obtained at different times in the same experiment (Figures 1a1c and 3a3b). Thus, individual ice crystals grow and fall through a cloud of supercooled droplets, moved by gravity and the air convection produced by the introduction of warm vapor into the base of the cloud chamber. Some of the ice crystals can begin to collect cloud droplets. As a consequence of all these dynamical processes taking place in the cloud chamber, it is likely that the crystals pass through a nonuniform spatial distribution of supercooled cloud droplets. Supplementary measurements performed during some particular runs show that the liquid water content may be reduced to half of its initial value one minute after ice crystal nucleation.

[29] The time elapsed between the ice crystal nucleation and the collection of crystals with droplet accretion was typically around 60 seconds. The observations suggest that the critical minimum crystal size necessary for the initial stage of capture and freezing of supercooled cloud droplets by hexagonal plates was around 60 μm diameter and for columnar ice crystals with width and length greater than 30 μm and 60 μm, respectively. However, we cannot state that these are the cutoff sizes of small ice crystals for riming because the data available so far are limited. At the moment it is possible to say that these are the smallest sizes that have been observed to collect.

[30] The sizes of ice crystals able to initiate the droplet collection process reported in the current work are smaller than those reported in previous studies. We conjecture that measurements in the field probably do not have a high chance of collecting ice crystals within a short period of time after their nucleation in clouds; particularly so in cases where the crystals are collected near the ground. Therefore, the ice crystals could grow further by vapor deposition after collecting the first droplet and before being sampled. Another possibility is that some of the authors have required snow crystals to accrete more than five cloud droplets to be classified as rimed in their data analysis [Ono, 1969; Harimaya, 1975]; surely, this would have extended the critical minimum sizes to larger crystal sizes. On the other hand, the theoretical studies do not consider the possible deformation of the flow fields around the ice crystal by the close proximity of a cloud droplet. Theoretical studies use the superposition method which assumes that each particle moves in an air-flow generated by the other particle falling in isolation. This method is very accurate when the relative fall velocity or separation of the particles are sufficiently large. In the case of ice particles around 60 μm diameter and cloud droplets around 20 μm diameter, the superposition method may not be appropriate. Instead, the Navier-Stokes equation needs to be solved by taking into account the border of both particles (boundary conditions), mainly when the particles are near each other.

[31] The observations of the riming characteristics of the ice crystals indicate that the droplets are collected preferentially around the border of the hexagonal plates. Some of the plates with droplet collection present a loss of symmetry (Figure 2), which is likely attributable to changes in the vapor density and temperature field around the crystal following droplet collection.

[32] No marked differences were observed between the cloud droplet size distribution and the size distribution of the droplets attached to ice crystals. In contrast with the theoretical predictions of Pitter and Pruppacher [1974] and Pitter [1977] and the observations of Harimaya [1975], we found cloud droplets with diameters less than 10 μm collected by snow crystals.

[33] In the present paper we report observations that ice crystals in clouds can start the riming process at smaller sizes than is reported in the literature, pointing out that the efficiency collision predictions of existing numerical models are not reliable for small ice crystals. These riming crystals are potentially important as graupel embryos in initiating precipitation processes. Furthermore, the shape and symmetry of snow crystals can be modified by the collection of a single cloud droplet, which could affect the subsequent growth by vapor deposition; thereby, complicated and irregular shapes of ice crystals can result. As is well known, the ice crystal size and shape distributions are fundamental parameters that determine the basic optical scattering properties of the ice clouds; particularly, cirrus clouds which can be the remnants of convective cloud.

[34] The surface structure of the ice crystals is an important parameter in the electric charge transferred in ice crystal/graupel collisions which is the main mechanism of cloud electrification [Saunders et al., 2001; Saunders, 2008]; this is another process where the early stage of riming can play a key role.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References

[35] We want to thank Peter Kelly and Paul Connolly for their assistance in this work. The Argentinian team thanks the SECYT, SECYT-UNC, CONICET and ANPCYT.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Mixed-Phase Cloud Chamber
  5. 3. Measurements and Results
  6. 4. Discussion and Conclusion
  7. Acknowledgments
  8. References
  • Finlayson-Pitts, B. J., and J. N. Pitts Jr. (2000), Chemistry of the Upper and Atmosphere, Academic, San Diego, Calif.
  • Harimaya, T. (1975), The riming properties of snow crystals, J. Meteorol. Soc. Jpn., 53, 384392.
  • Kajikawa, M. (1974), On the collection efficiency of snow crystals for cloud droplets, J. Meteorol. Soc. Jpn., 52, 328335.
  • MacGorman, D. R., and W. D. Rust (1998), The Electrical Nature of Storms, Oxford Univ. Press, New York.
  • Ono, A. (1969), The shape and riming properties of ice crystals in natural clouds, J. Atmos. Sci., 26, 138147.
  • Pitter, R. L. (1977), A reexamination of riming on thin ice plates, J. Atmos. Sci., 34, 684685.
  • Pitter, R. L., and H. R. Pruppacher (1974), A numerical investigation of collision efficiencies of simple ice plates colliding with supercooled drops, J. Atmos. Sci., 31, 551559.
  • Pruppacher, H. R., and J. D. Klett (1997), Microphysics of Clouds and Precipitation, Atmos. Oceanogr. Sci. Libr., vol. 18, 2nd ed., 954 pp., Kluwer Acad., Dordrecht, Netherlands.
  • Reinking, R. (1979), The onset and steady growth of snow crystals by accretion of droplets, J. Atmos. Sci., 36, 870881.
  • Saunders, C. P. R. (2008), Charge separation mechanisms in clouds, Space Sci. Rev., 137, 335353.
  • Saunders, C. P. R., S. L. Peck, G. G. Aguirre Varela, E. Avila, and N. Castellano (2001), A laboratory study of the influence of the water vapour and mixing on the charge transfer process during collisions between ice crystals and graupel, Atmos. Res., 58, 187203.
  • Schaefer, V. J. (1956), The preparation of snow crystal replicas-VI, Weatherwise, 9, 132135.
  • Schlamp, R. J., H. R. Pruppacher, and A. E. Hamielec (1975), A numerical investigation of the efficiency with which simple columnar ice crystals collide with supercooled water drops, J. Atmos. Sci., 32, 23302337.
  • Wang, P. K., and W. Ji (2000), Collision efficiencies of ice crystals at low-intermediate Reynolds numbers colliding with supercooled cloud droplets: A numerical study, J. Atmos. Sci., 57, 10011009.
  • Young, K. C. (1993), Microphysical Processes in Clouds, Oxford Univ. Press, New York.