• Bailey, M., , and J. Hallett, 2002: Nucleation effects on the habit of vapour grown ice crystals from −18° to −42° C. Quart. J. Roy. Meteor. Soc., 128 , 14611483.

    • Search Google Scholar
    • Export Citation
  • Bailey, M., , and J. Hallett, 2004: Growth rates and habits of ice crystals between −20° and −70° C. J. Atmos. Sci., 61 , 514544.

  • Bentley, W. A., , and W. J. Humphreys, 1931: Snow Crystals. McGraw-Hill, 227 pp.

  • Evans, A. G., , J. D. Locatelli, , M. T. Stoelinga, , and P. V. Hobbs, 2005: The IMPROVE-1 storm of 1–2 February 2001. Part II: Cloud structures and the growth of precipitation. J. Atmos. Sci., 62 , 34563473.

    • Search Google Scholar
    • Export Citation
  • Gordon, G. L., , and J. D. Marwitz, 1986: Hydrometeor evolution in rainbands over the California valley. J. Atmos. Sci., 43 , 10871100.

  • Heymsfield, A. J., , and L. M. Miloshevich, 2003: Parameterizations for the cross-sectional area and extinction of cirrus and stratiform ice cloud particles. J. Atmos. Sci., 60 , 936956.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., , S. Lewis, , A. Bansemer, , J. Iaquinta, , L. M. Miloshevich, , M. Kajikawa, , C. Twohy, , and M. R. Poellot, 2002: A general approach for deriving the properties of cirrus and stratiform ice cloud particles. J. Atmos. Sci., 59 , 329.

    • Search Google Scholar
    • Export Citation
  • Houze Jr., R. A., , P. V. Hobbs, , P. H. Herzegh, , and D. B. Parsons, 1979: Size distributions of precipitation particles in frontal clouds. J. Atmos. Sci., 36 , 156162.

    • Search Google Scholar
    • Export Citation
  • Knollenberg, R. G., 1976: Three new instruments for cloud physics measurements: The 2-D spectrometer, the forward scattering spectrometer probe, and the active scattering aerosol spectrometer. Preprints, Int. Conf. on Cloud Physics, Boulder, CO, Amer. Meteor. Soc., 554–561.

  • Korolev, A., , and B. Sussman, 2000: A technique for habit classification of cloud particles. J. Atmos. Oceanic Technol., 17 , 10481057.

    • Search Google Scholar
    • Export Citation
  • Korolev, A., , and G. Isaac, 2003: Roundness and aspect ratio of particles in ice clouds. J. Atmos. Sci., 60 , 17951808.

  • Korolev, A., , G. A. Isaac, , and J. Hallett, 1999: Ice particle habits in Arctic clouds. Geophys. Res. Lett., 26 , 12991302.

  • Korolev, A., , G. A. Isaac, , and J. Hallett, 2000: Ice particle habits in stratiform clouds. Quart. J. Roy. Meteor. Soc., 126 , 28732902.

    • Search Google Scholar
    • Export Citation
  • Lawson, P. R., , B. A. Baker, , C. G. Schmitt, , and T. L. Jensen, 2001: An overview of microphysical properties of Arctic clouds observed in May and July 1998 during FIRE ACE. J. Geophys. Res., 106 , 1498915014.

    • Search Google Scholar
    • Export Citation
  • Lo, K. K., , and R. E. Passarelli Jr., 1982: The growth of snow in winter storms: An airborne observational study. J. Atmos. Sci., 39 , 697706.

    • Search Google Scholar
    • Export Citation
  • Magono, C., , and C. Lee, 1966: Meteorological classification of natural snow crystals. J. Fac. Sci. Hokkaido Univ. Ser. 7, 2 , 321335.

  • Nakaya, U., 1954: Snow Crystals, Natural and Artificial. Harvard University Press, 510 pp.

  • Stoelinga, M. T., and Coauthors, 2003: Improvement of microphysical parameterization through observational verification experiment. Bull. Amer. Meteor. Soc., 84 , 18071826.

    • Search Google Scholar
    • Export Citation
  • Vardiman, L., , and C. L. Hartzell, 1976: Final report on an investigation of precipitating ice crystals from natural and seeded winter orographic clouds. Bureau of Reclamation Tech. Rep. SR-359-47, U.S. Dept. of Interior, Denver, CO, 129 pp. [Available from the National Technical Information Service, 5285 Port Royal Rd., Springfield, VA 22161.].

  • Walden, V. P., , S. G. Warren, , and E. Tuttle, 2003: Atmospheric ice crystals over the Antarctic Plateau in winter. J. Appl. Meteor., 42 , 13911405.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Topographic map of Washington and Oregon showing locations of the three snow particle observing sites. Elevation shading scale is shown in upper lhs.

  • View in gallery

    Images of snow crystals viewed through a microscope by ground-based snow observers during the IMPROVE-2 field study: (top) a dendritic crystal with plates at ends (Magono and Lee’s type P2c), with a few frozen droplets attached; (middle) an unrimed bullet rosette (Magono and Lee’s type C2a); and (bottom) a moderately rimed dendritic crystal with plates at ends (Magono and Lee’s type P2c). A 1-mm scale in upper lhs applies to all images.

  • View in gallery

    Pictures of Magono and Lee’s irregular snow crystal types, adapted from Magono and Lee (1966): (a) ice particles, (b) rimed particles, (c) broken branch, (d) rimed broken branch, (e) rimed broken branch, and (f) miscellaneous.

  • View in gallery

    Surface observations of snow crystal types and precipitation rates at Corbett Park, Santiam Pass, and Tombstone Pass, Oregon, on 4–5 Dec 2001. See Fig. 5 for snow-shape symbols and Fig. 1 for observation locations.

  • View in gallery

    Classification and symbols for snow particles indicated in Fig. 4.

  • View in gallery

    Image strips from an aircraft-mounted 2D-C particle imaging probe: (a) images gathered upwind of Santiam Pass, Oregon, at 0050 UTC 5 Dec 2001 at a temperature of −19°C; (b) same as in (a), but at 0127 UTC at −11°C; (c) 2D-C images simulated from silhouettes of idealized cold-type snow crystals shown in Fig. 7; and (d) examples of images classified as irregular from Korolev et al. (2000). The vertical width of all strips is 800 μm.

  • View in gallery

    Idealized cold-type snow crystals from Magono and Lee (1966, their Fig. 1): (a) radiating assemblage of plates; (b) bullets with plates; (c) sideplanes; (d) scalelike sideplanes; and (e) combination of sideplanes, bullets, and columns.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 54 54 8
PDF Downloads 35 35 1

The Occurrence of “Irregular” Ice Particles in Stratiform Clouds

View More View Less
  • 1 Department of Atmospheric Sciences, University of Washington, Seattle, Washington
© Get Permissions
Full access

Abstract

Recent studies that have classified ice particles from airborne imaging probe data have concluded that the vast majority of ice particles in stratiform precipitation systems are of an “irregular shape.” This conclusion stands in contrast to the findings from microscope observations of snow particles at the ground during the Improvement of Microphysical Parameterization through Observational Verification Experiment from November to December 2001 in the Oregon Cascade Mountains (IMPROVE-2), which show that most snow crystals (either single crystals or the component crystals of snow aggregates) are readily identified as regular types within established crystal classification systems. This apparent contradiction is rectified by examining the definition of the term irregular as applied to ice particles and by considering limitations of different methods for observing ice particles. It is concluded that the finding of the airborne probe-based studies is a consequence of both limitations of the observing technology and an overly broad definition of irregular shape that is not consistent with the more restrictive definition established in well-known snow crystal classification schemes. When detailed microscope analysis of snow crystals is performed at the ground, and all regular types are included in the classification, the vast majority of snow crystals are of an identifiable regular type, rather than an irregular type.

The classification of the vast majority of particles as irregular implies that there is little hope to describe the important properties of these particles (such as their scattering properties, fall speeds, and temperature and humidity conditions in which they grew), when in fact, many of these particles are of known types with known properties. Instead of using the term irregular, classification studies should use a term that focuses on the limitation of the observation method as being the defining characteristic of the category, such as “unidentified” or “undetermined.”

Corresponding author address: Mark T. Stoelinga, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195-1640. Email: stoeling@atmos.washington.edu

Abstract

Recent studies that have classified ice particles from airborne imaging probe data have concluded that the vast majority of ice particles in stratiform precipitation systems are of an “irregular shape.” This conclusion stands in contrast to the findings from microscope observations of snow particles at the ground during the Improvement of Microphysical Parameterization through Observational Verification Experiment from November to December 2001 in the Oregon Cascade Mountains (IMPROVE-2), which show that most snow crystals (either single crystals or the component crystals of snow aggregates) are readily identified as regular types within established crystal classification systems. This apparent contradiction is rectified by examining the definition of the term irregular as applied to ice particles and by considering limitations of different methods for observing ice particles. It is concluded that the finding of the airborne probe-based studies is a consequence of both limitations of the observing technology and an overly broad definition of irregular shape that is not consistent with the more restrictive definition established in well-known snow crystal classification schemes. When detailed microscope analysis of snow crystals is performed at the ground, and all regular types are included in the classification, the vast majority of snow crystals are of an identifiable regular type, rather than an irregular type.

The classification of the vast majority of particles as irregular implies that there is little hope to describe the important properties of these particles (such as their scattering properties, fall speeds, and temperature and humidity conditions in which they grew), when in fact, many of these particles are of known types with known properties. Instead of using the term irregular, classification studies should use a term that focuses on the limitation of the observation method as being the defining characteristic of the category, such as “unidentified” or “undetermined.”

Corresponding author address: Mark T. Stoelinga, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195-1640. Email: stoeling@atmos.washington.edu

1. Introduction

A series of papers (Korolev et al. 1999, 2000; Korolev and Isaac 2003) described results from a manual and automated particle identification analysis of two-dimensional imagery collected by aircraft flying through stratiform precipitation systems. These studies concluded that 80%–97% of particles in stratiform clouds, at temperatures ranging from 0° to −40°C, are of an “irregular” shape. Korolev et al. (1999, 2000) considered this result to be significant because it contradicted the general impression left by published photographic collections of snow crystals observed at the ground (e.g., Bentley and Humphreys 1931; Nakaya 1954) that irregular crystals do not occur very often. They hypothesized that such photographic collections were biased toward showing mostly examples of pristine monocrystals (such as dendrites, stellars, plates, and columns) because of their aesthetic appearance, leaving an impression that most snow crystals are pristine monocrystals.

The conclusion that the most naturally occurring ice particles are of an irregular shape seems to contradict our experience that most snow crystals (either single crystals or the component crystals of snow aggregates) are readily identified as regular types within established crystal classification systems, based on careful stereo microscope observations of snow particles at the ground during the Improvement of Microphysical Parameterization through Observational Verification Experiment from November to December 2001 in the Oregon Cascade Mountains (IMPROVE-2). In this note, we attempt to rectify this apparent contradiction by examining the definition of the term irregular as applied to naturally occurring snow crystals and by considering limitations of different methods for observing naturally occurring snow crystals.

2. Irregular snow crystals

It is well known that natural snow crystals take on a wide variety of shapes and structures. The most basic “building blocks” of crystal growth are the habits that occur when growth occurs predominantly on the basal faces, the prism faces, or the corners between the prism faces of an ice crystal. These three growth regimes result in columnar/needle-type growth, hexagonal plate-type growth, and stellar/dendritic growth, respectively. The growth regime is determined by local conditions of temperature and relative humidity with respect to both liquid water and ice (Magono and Lee 1966; Bailey and Hallett 2002, 2004). If a particle grows for some time under one set of conditions, it typically forms a single-habit particle, such as a column, hexagonal plate, needle, bullet, dendrite, stellar, etc. If a particle of one type moves into a region where a different growth regime is favored, it can take on a multihabit form, such as a capped column (a column capped with hexagonal plates) or a plate with dendritic extensions.

Particles can also be classified as “monocrystals” and “polycrystals.” Monocrystals are characterized by approximately hexagonally symmetric growth about a single c axis (the c axis is the principal axis, normal to the basal face of the hexagonal lattice). Polycrystals form when imperfections result in the growth of multiple parts, often emanating from a central point such as a frozen droplet, each having its own c axis that is randomly oriented relative to the c axes of the other parts. Both monocrystals and polycrystals may be single-habit types (e.g., monocrystalline plates, needles, columns, dendrites, bullets, etc., or polycrystalline bullet rosettes, radiating assemblages of dendrites, etc.), or they may be multihabit types (e.g., monocrystalline capped columns, dendrites with plates at ends, etc., or polycrystalline plates with spatial dendrites, bullet rosettes with plates, etc.). [See Magono and Lee (1966) for sample illustrations of these and many other particle types.]

If a crystal grows essentially undisturbed in any habit, it is often referred to as “pristine.” However, several processes can modify the structure and appearance of falling crystals to varying degrees. They can collect supercooled cloud droplets (forming rimed particles and, eventually, graupel), they can break into fragments, and they can partially sublimate in ice-subsaturated conditions. Crystals also collide and stick to each other, forming aggregates. In our experience, most snow particles larger than ∼2 mm in diameter are actually aggregates of several crystals. However, our focus here is on the shapes of single crystals, whether falling as a single particle or as a component of an aggregate. It should be noted that even with a significant degree of modification, the original crystal type can often be determined under close inspection with a microscope.

Elaborate classification schemes for snow crystals have been developed by Nakaya (1954) and Magono and Lee (1966). Nakaya’s scheme included 41 different types. Magono and Lee extended Nakaya’s scheme to 81 types, finding that while Nakaya’s scheme worked well for pristine monocrystals, it was too simple for modified and polycrystalline particles. Both Nakaya (1954) and Magono and Lee (1966) recognized that some particles did not fit any of their defined types, referring to these particles as irregular. Although their descriptions of irregular crystals were slightly different, they both included in this category particles that had undergone a complex history of growth and/or modification such that even under careful microscope inspection, a distinct particle type was difficult to ascertain. However, it is important to recognize that Nakaya’s and Magono and Lee’s irregular types were narrowly defined because their elaborate classification schemes (particularly that of Magono and Lee) covered the full variety of multihabit crystals and polycrystals that were identifiable under careful microscope observation. It is quite clear that neither Nakaya nor Magono and Lee automatically included polycrystals, multihabit crystals, or modified crystals in their irregular category, unless a distinct particle type was not identifiable under microscope observation. Magono and Lee’s classification scheme is now widely used and referenced in the literature. A more recent scheme put forward by Bailey and Hallett (2004), while organizing the regular types somewhat differently than Magono and Lee, essentially affirms Magono and Lee’s definition of the irregular type.

On the other hand, the definition of irregular particles used by Korolev et al. (1999, 2000) is much broader than that established by Nakaya and Magono and Lee. Korolev et al. (1999) performed manual classification of images gathered by a Cloud Particle Imager (CPI; Lawson et al. 2001) and stated that “particles were categorized into two groups: pristine and irregular. The term “pristine ice” is applied to faceted ice single crystals. The rest of the particles fall into the irregular category.” They also described irregular crystals as consisting mostly of either polycrystals or partially sublimated crystals. Thus, a succinct interpretation of Korolev et al.’s (1999) definition is that irregulars are anything other than pristine monocrystals.

Korolev et al. (2000) performed an automated classification of particle images recorded with a 2D-C probe (Knollenberg 1976), using the technique of Korolev and Sussman (2000). This method attempted to identify “spheres” (particles with circular images), “needles” (particles that are at least 3 times longer than they are wide), and “dendrites” (dendrites, stellar crystals, and aggregates of dendrites). [For a complete description of these definitions, see Korolev et al. (2000).] Korolev et al. (2000) described irregulars as follows: “Polycrystalline ice particles, such as combinations of plates and columns, heavily rimed particles, and graupel, and other forms that do not display the features of needles, dendrites, or spheres would fall into this category.” Though differing slightly from Korolev et al. (1999), Korolev et al. (2000) again essentially defined irregulars as anything other than pristine monocrystals.

Korolev and Isaac (2003) performed an automated analysis of CPI imagery of ice particles. While they did not classify particle types, they examined the roundness and aspect ratio of particles, and they found that their results supported the conclusion of Korolev et al. (1999, 2000) that in stratiform clouds, “the majority of ice particles fall into an irregular category.” However, the meaning of this conclusion was broadened by a significant statement in their introduction: “A simple visual analysis of irregularly shaped particles shows a great variety of different forms, which cannot be covered by any of the existing classifications of cloud ice particles (e.g., Magono and Lee 1966).” The implication of this statement is that the vast majority of ice particles in stratiform clouds do not conform to any (not just pristine monocrystalline) types in Magono and Lee (1966). Whether or not Korolev and Isaac (2003) meant to imply such a sweeping conclusion is not clear. In any event, some clarification of the frequency of occurrence of irregular ice particles is in order.

3. Snow crystal observations during IMPROVE-2

The Improvement of Microphysical Parameterization through Observational Verification Experiment was conducted from January to February 2001 on the coast of Washington State (IMPROVE-1) and from November to December 2001 in the Oregon Cascade Mountains (IMPROVE-2). The purpose of these field projects was to improve the representations of cloud and precipitation processes in mesoscale models (Stoelinga et al. 2003). During the 4 weeks of the IMPROVE-2 field project, falling snow particles were observed at the ground with multipower stereo microscopes at three locations near the crest of the Oregon Cascade Mountains in the vicinity of Santiam Pass (see Fig. 1 for locations). Observations were made every 15 min, usually at all 3 sites, but occasionally only at 1 or 2. The use of a stereo microscope provides a sharpness, depth perception, and definition of internal structural detail that facilitates the identification of most particle shapes, from simple plates and columns to highly three-dimensional polycrystalline particles such as spatial dendrites or radiating assemblages of plates or sectors. Sample images of some crystals (1–3-mm maximum dimension) observed during our study are shown in Fig. 2. The magnification used (40x) allowed for a large field of view (∼4.5 mm), such that larger particles could be seen in their entirety, or a number of smaller particles at once. At the same time, details at the scale of ∼10 μm could be seen by the observers, looking directly in the microscope. The sharpness of detail is somewhat less in the photographic images, due to limitations of the camera focus, lighting, exposure, and digitization. Other studies using more elaborate and high-quality microphotography have presented sharper images of crystals viewed under a microscope (Magono and Lee 1966; Heymsfield et al. 2002; Walden et al. 2003). In particular, the Walden et al. study shows images of distinctly hexagonal plates that are only 15 μm in diameter.

Snow particles were collected on glass slides and then immediately examined under the microscope. Both individual crystals and components of aggregates were examined. While the observers did not precisely count and then identify every crystal that was collected on the slides, they did visually scan the entire slide, and recorded in notebooks any crystal shapes observed according to the classification scheme of Magono and Lee (1966), with qualitative frequency of occurrence of each type noted. Degree of riming was also noted. The number of individual crystals on a slide would vary from tens to hundreds, depending on the precipitation rate. The observers were keenly aware of the tendency for attention to be preferentially drawn toward the larger and more pristine crystals and therefore were careful to examine a representative sampling of the full size distribution on the slide. The impact of the falling crystals on the slide was gentle enough that fragmentation was not a problem. Fragments were observed, but were isolated on the slide, indicating that they were already fragmented before impact. The process of visually identifying the habit types was usually fast enough such that sublimation and melting did not preclude proper identification. If melting precluded the proper identification of habit types, observations were discontinued. Other meteorological observations, such as snow rate (−SN, SN, and +SN for light, moderate, and heavy snowfall rates) and wind conditions, were also recorded. The temperatures at the observation site were in the range from approximately −5° to +1°C when observations were being taken. In spite of this confined temperature range, a wide variety of particle types were encountered because of the variability of the altitude of the lifting region aloft under variable synoptic conditions, ranging from deep frontal stratiform precipitation to shallow, postfrontal orographic precipitation.

The full 4 weeks of the snow crystal dataset contained observations of crystals on 854 individual glass slides. The observers nearly always found that as long as the crystals were not significantly melted or rimed, they were able to identify the shapes of all the crystals on the slides as some regular type in Magono and Lee’s classification scheme. There were occasions when particles were classified as one of Magono and Lee’s five irregular types (illustrated in Fig. 3): 47 occasions when “broken branches”1 of either dendrites or sectors were observed with various degrees of riming and 25 occasions when either “rimed particles” or “ice particles” (unidentifiable because of modification or small size) were observed. These irregular particles were usually seen in conjunction with other identifiable crystals. It is estimated that <5% of the total number of particles observed were of Magono and Lee’s irregular type.

An example of a 6.5-h time series of snow crystal observations, taken on 4–5 December 2001, is shown in Fig. 4. All crystals identified as being present at the time of the observation were plotted, with the exception of crystals that were recorded as few in number. An explanation of the symbols for the different snow particle types is given in Fig. 5. This series of observations occurred during the passage of a deep, stratiform frontal precipitation band, in conjunction with upslope flow at all levels on the west slope of the Cascade Range. Because riming of the snow crystals was mostly moderate or less, most of the crystals were readily identifiable as regular types in the Magono and Lee scheme. As on several other days during IMPROVE-2, the three snow particle observers stationed at different sites independently identified many of the same types. During this precipitation event, the observers found crystal types representing a wide range of diffusional growth temperatures including cold types (temperature <−20°C), such as assemblages of plates, sideplanes, and bullets, and warm types (temperature >−20°C), such as plates with sectorlike extensions, plates with dendritic extensions, dendrites, and capped columns. Irregular types (broken branches, rimed particles, and ice particles) were reported in 13 of the 62 observations during this event and are estimated to have composed <8% of the total number of crystals observed. The irregular percentages did not differ significantly among the three observers.

This method of observation has an inherently subjective component to it because the data are composed of notes written by human observers scanning a microscope slide and recording their subjective determination of the crystal types seen. However, this alone cannot explain the nearly opposite results of our microscope observations at the ground (finding that <10% of snow crystals are of an irregular shape) and those of classification schemes applied to imagery from airborne probes (finding that 80%–97% of ice particles are of an irregular shape). In the next section, we examine this discrepancy by comparing ground and airborne observations of snow particles in an IMPROVE-2 case study.

4. Comparison of IMPROVE-2 ground observations with aircraft particle imagery

During IMPROVE-2, while observers at the ground were examining the crystal types under a microscope, the University of Washington Convair-580 research aircraft was gathering cloud microphysical measurements aloft and upwind of the ground sites, including images of ice particles using a 2D-C probe with 25-μm resolution. Samples of particle images collected by the aircraft at temperatures of −19° and −11°C are shown in Figs. 6a,b, respectively. It is clearly difficult to ascertain by eye the types of any of the particles that appear in the image strips at −19°C (Fig. 6a). In the image strips at −11°C (Fig. 6b), several dendritic-type particles appear, and a few plates, but there are also many particles that are unidentifiable. The Korolev and Sussman (2000) algorithm was applied to these different temperature regions, along flight segments of 3-min duration at −19°C and 4-min duration at −11°C, comprising several thousand 2D-C particle images each. The algorithm found that 88% of the particles seen at −19°C were irregular, with the small remainder approximately evenly distributed among spherical, needle, and dendritic; at −11°C, the algorithm found mostly irregular and dendritic (in approximately equal portions), with a few spherical or needle particles.

However, it is quite reasonable to assume that the particles seen in these images are of the same types as those seen at the ground downwind of the aircraft. The precipitation was fairly steady for an extended period of time, and estimates of particle trajectories place the aircraft near the trajectories of particles falling at the observing sites. Furthermore, it is difficult to envision that the particles would become more pristine as they fell between the aircraft and the ground. Therefore, it is likely that the image strips at −19°C include the cold-type crystals such as assemblages of plates, sideplanes, and capped bullets seen on the ground at the 3 sites. At −11°C, the images likely include both the cold-type crystals and warm-type crystals (plates with sectorlike extensions, plates with dendritic extensions, dendrites, capped columns, and stellars) that were observed at the ground. Thus, the unidentifiable particles that compose most of the images at −19°C and a significant fraction of the images at −11°C are likely the cold-type crystals observed at the ground that conform to regular types in the Magono and Lee (1966) classification scheme.

The plausibility of this assumption can be illustrated by creating a strip of simulated 2D-C images of cold-type crystals, shown in Fig. 6c. This strip was constructed from idealized drawings of cold-type polycrystals from the Magono and Lee (1966) classification chart, shown in Fig. 7. To simulate realistic 2D images, the drawings were silhouetted, and then a random sample of the silhouettes was selected at different sizes (ranging from 500 to 1000 μm) and orientations. The edges of the silhouettes were slightly roughened, consistent with the degree of roughness seen on the edges of pristine crystals in real 2D-C images. Finally, the images were pixilated at 25-μm resolution. It can be seen that the resulting particle images (Fig. 6c) become visually unidentifiable in terms of crystal type and look similar to many of the unidentifiable particles in the actual 2D-C images collected during IMPROVE-2 (Figs. 6a,b), as well as the examples of particles identified as irregular in Korolev et al. (2000; Fig. 6d).

5. Limitations of 2D classification schemes

A drawback of classification methods that rely on airborne particle imagery is that the ability to discern the 3D structure of particles is limited by a 2D view of finite resolution. Classification schemes are generally confined to a few specific shapes that are clearly identifiable from the imagery, resulting in a large class of irregular particles that do not fit those shapes. Furthermore, as particle size decreases, the number of particles present increases exponentially. For a typical slope parameter λ = 2500 m−1 for an exponential ice particle size distribution (e.g., Houze et al. 1979; Lo and Passarelli 1982; Gordon and Marwitz 1986), there are 6 times as many 80-μm particles as there are 800-μm particles. Finite image resolution increases the likelihood that the smaller particles will not match any of the specified shapes, as corners become rounded and details become obscured. This potentially results in a large bias toward irregular particles.

There is some evidence of this issue in Korolev et al. (2000), who used a large dataset of images from a 2D-C probe, examining the dependence of classifiability on particle size. They found that 76% of the 1.6 million particles larger than 500 μm were classified as irregular, whereas 88% of the 3.6 million particles between 125 and 500 μm were classified as irregular. This issue is not confined to airborne studies. A ground-based study in Colorado by Vardiman and Hartzell (1976) classified particles as either plates, columns, dendrites, graupel, or irregular, using images from an instrument that automatically photographed ice particles collected on a continuously moving belt. The photos provided an effective resolution that is probably similar to that of the 2D-C instrument. They found irregulars to occur 73% of the time. However, their irregular particles had an average diameter of 472 μm, a size at which little detail can be discerned from the photos, whereas the particles classified as dendrites had an average diameter of 1577 μm. These studies suggest that the ability to define a particle’s type is significantly influenced by its size.

Korolev and Isaac (2003) found that the average roundness and aspect ratio of particles imaged with a CPI are relatively constant for particles larger than ∼90 μm, but increase rapidly (toward circular values) for particles smaller than that size. They describe such particles as “mostly irregular in shape.” Heymsfield and Miloshevich (2003) found a similar result in terms of rapidly increasing area ratio (similar to roundness) with decreasing diameter for particles smaller than 100 μm, using a balloonborne replicator with higher image clarity than the CPI. However, they did not describe such particles as mostly irregular, and the particles appear mostly platelike in the images presented. Indeed, as Korolev et al. (2000) point out, there are physical reasons to expect habit-type differences for different size ranges (e.g., dendritic features generally do not appear until a particle is 400–600 μm in size). In any event, the large number of more round-looking particles at small sizes, whether due to actual habit characteristics or imaging resolution problems, does not imply prevalence of an irregular shape in either case.

Certainly modern instruments such as the CPI represent a significant improvement over previous particle imaging techniques, with an order-of-magnitude smaller pixel size than the 2D-C (2.5 versus 25 μm) and better depiction of internal structure. However, the CPI is still limited to a two-dimensional view in a single focus plane. In contrast, manual stereo microscope observations allow for the examination of particles in a 3D view and in multiple focus planes, with perimeter and internal details discernible at a scale equal to or better than that of CPI imagery. Many well-formed polycrystals are difficult to identify without these unique capabilities of the stereo microscope. It should also be noted, however, that even with the CPI, careful manual classification can result in the identification of a significant number of regular-habit particles that existing automated classification schemes would likely identify as irregular (e.g., Evans et al. 2005).

6. Conclusions

This note has examined the conclusion from aircraft-based studies of ice particles in stratiform clouds (Korolev et al. 1999, 2000; Korolev and Isaac 2003) that the vast majority of particles are of an “irregular” shape, a conclusion that contradicts our experience with careful stereo microscope observations of snow particles at the ground. We have found that most snow crystals (either single crystals or the component crystals of snow aggregates) are readily identifiable as regular types within established snow crystal classification schemes. We examined this contradiction from several perspectives, including a review of the definitions of irregular ice particles, a consideration of the limitations of aircraft particle imaging instruments compared to manual microscope observations, and a case study comparing aircraft particle imagery and ground-based microscope observations of snow particles in the Oregon Cascade Mountains during the IMPROVE-2 field study. We conclude that the seemingly contradictory results arise from

  1. a very broad definition of irregular shape that is used in the aircraft-probe-based studies, compared to the more restricted definition established in well-known snow crystal classification schemes;
  2. the difficulty of classifying particle types based on 2D images of particles with limited resolution and depth of field [as discussed in Korolev et al. (2000)], compared to stereo microscope observations;
  3. the exacerbation of issue 2) as particle diameter decreases, since image clarity decreases with particle size but number concentration increases exponentially.

While we generally agree with Korolev and collaborators that few naturally occurring snow crystals are of the pristine monocrystalline type that are disproportionately represented in photographic collections of snow crystals, we believe their conclusion that the vast majority of ice particles are irregular arises from a classification scheme that is based largely upon what can or cannot be discerned from images of finite resolution, rather than upon the true physical structure of the particles observed. An unfortunate consequence of this conclusion is the implication that there is little hope to describe the important microphysical and radiative properties of these irregular particles, when in fact, many of these particles are of known types with known properties. Instead of using the term irregular, such studies should use a term that focuses on the limitation of the observation method as being the defining characteristic of the category, such as “unidentified” or “undetermined.” More extensive studies that use a combination of airborne probes, ground observations, and remotely sensed measurements may allow for an integration of information that creates a clearer picture of the true occurrence of the many regular ice crystal types in stratiform precipitation systems. Such information is invaluable for understanding the history of conditions in which particles grew, their fall speed and precipitation rate, and their radiative properties.

Acknowledgments

This research was supported by Grants ATM-9908446 and ATM-0242592 from the National Science Foundation. We wish to thank Matt Garvert (University of Washington) and Greg Thompson (NCAR), who spent many hours in the cold (along with the second author) collecting observations of snow particles during IMPROVE-2. We also thank Arthur Rangno for fruitful discussions and Thomas Wilson for assistance with the analysis of aircraft data.

REFERENCES

  • Bailey, M., , and J. Hallett, 2002: Nucleation effects on the habit of vapour grown ice crystals from −18° to −42° C. Quart. J. Roy. Meteor. Soc., 128 , 14611483.

    • Search Google Scholar
    • Export Citation
  • Bailey, M., , and J. Hallett, 2004: Growth rates and habits of ice crystals between −20° and −70° C. J. Atmos. Sci., 61 , 514544.

  • Bentley, W. A., , and W. J. Humphreys, 1931: Snow Crystals. McGraw-Hill, 227 pp.

  • Evans, A. G., , J. D. Locatelli, , M. T. Stoelinga, , and P. V. Hobbs, 2005: The IMPROVE-1 storm of 1–2 February 2001. Part II: Cloud structures and the growth of precipitation. J. Atmos. Sci., 62 , 34563473.

    • Search Google Scholar
    • Export Citation
  • Gordon, G. L., , and J. D. Marwitz, 1986: Hydrometeor evolution in rainbands over the California valley. J. Atmos. Sci., 43 , 10871100.

  • Heymsfield, A. J., , and L. M. Miloshevich, 2003: Parameterizations for the cross-sectional area and extinction of cirrus and stratiform ice cloud particles. J. Atmos. Sci., 60 , 936956.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., , S. Lewis, , A. Bansemer, , J. Iaquinta, , L. M. Miloshevich, , M. Kajikawa, , C. Twohy, , and M. R. Poellot, 2002: A general approach for deriving the properties of cirrus and stratiform ice cloud particles. J. Atmos. Sci., 59 , 329.

    • Search Google Scholar
    • Export Citation
  • Houze Jr., R. A., , P. V. Hobbs, , P. H. Herzegh, , and D. B. Parsons, 1979: Size distributions of precipitation particles in frontal clouds. J. Atmos. Sci., 36 , 156162.

    • Search Google Scholar
    • Export Citation
  • Knollenberg, R. G., 1976: Three new instruments for cloud physics measurements: The 2-D spectrometer, the forward scattering spectrometer probe, and the active scattering aerosol spectrometer. Preprints, Int. Conf. on Cloud Physics, Boulder, CO, Amer. Meteor. Soc., 554–561.

  • Korolev, A., , and B. Sussman, 2000: A technique for habit classification of cloud particles. J. Atmos. Oceanic Technol., 17 , 10481057.

    • Search Google Scholar
    • Export Citation
  • Korolev, A., , and G. Isaac, 2003: Roundness and aspect ratio of particles in ice clouds. J. Atmos. Sci., 60 , 17951808.

  • Korolev, A., , G. A. Isaac, , and J. Hallett, 1999: Ice particle habits in Arctic clouds. Geophys. Res. Lett., 26 , 12991302.

  • Korolev, A., , G. A. Isaac, , and J. Hallett, 2000: Ice particle habits in stratiform clouds. Quart. J. Roy. Meteor. Soc., 126 , 28732902.

    • Search Google Scholar
    • Export Citation
  • Lawson, P. R., , B. A. Baker, , C. G. Schmitt, , and T. L. Jensen, 2001: An overview of microphysical properties of Arctic clouds observed in May and July 1998 during FIRE ACE. J. Geophys. Res., 106 , 1498915014.

    • Search Google Scholar
    • Export Citation
  • Lo, K. K., , and R. E. Passarelli Jr., 1982: The growth of snow in winter storms: An airborne observational study. J. Atmos. Sci., 39 , 697706.

    • Search Google Scholar
    • Export Citation
  • Magono, C., , and C. Lee, 1966: Meteorological classification of natural snow crystals. J. Fac. Sci. Hokkaido Univ. Ser. 7, 2 , 321335.

  • Nakaya, U., 1954: Snow Crystals, Natural and Artificial. Harvard University Press, 510 pp.

  • Stoelinga, M. T., and Coauthors, 2003: Improvement of microphysical parameterization through observational verification experiment. Bull. Amer. Meteor. Soc., 84 , 18071826.

    • Search Google Scholar
    • Export Citation
  • Vardiman, L., , and C. L. Hartzell, 1976: Final report on an investigation of precipitating ice crystals from natural and seeded winter orographic clouds. Bureau of Reclamation Tech. Rep. SR-359-47, U.S. Dept. of Interior, Denver, CO, 129 pp. [Available from the National Technical Information Service, 5285 Port Royal Rd., Springfield, VA 22161.].

  • Walden, V. P., , S. G. Warren, , and E. Tuttle, 2003: Atmospheric ice crystals over the Antarctic Plateau in winter. J. Appl. Meteor., 42 , 13911405.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Topographic map of Washington and Oregon showing locations of the three snow particle observing sites. Elevation shading scale is shown in upper lhs.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

Fig. 2.
Fig. 2.

Images of snow crystals viewed through a microscope by ground-based snow observers during the IMPROVE-2 field study: (top) a dendritic crystal with plates at ends (Magono and Lee’s type P2c), with a few frozen droplets attached; (middle) an unrimed bullet rosette (Magono and Lee’s type C2a); and (bottom) a moderately rimed dendritic crystal with plates at ends (Magono and Lee’s type P2c). A 1-mm scale in upper lhs applies to all images.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

Fig. 3.
Fig. 3.

Pictures of Magono and Lee’s irregular snow crystal types, adapted from Magono and Lee (1966): (a) ice particles, (b) rimed particles, (c) broken branch, (d) rimed broken branch, (e) rimed broken branch, and (f) miscellaneous.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

Fig. 4.
Fig. 4.

Surface observations of snow crystal types and precipitation rates at Corbett Park, Santiam Pass, and Tombstone Pass, Oregon, on 4–5 Dec 2001. See Fig. 5 for snow-shape symbols and Fig. 1 for observation locations.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

Fig. 5.
Fig. 5.

Classification and symbols for snow particles indicated in Fig. 4.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

Fig. 6.
Fig. 6.

Image strips from an aircraft-mounted 2D-C particle imaging probe: (a) images gathered upwind of Santiam Pass, Oregon, at 0050 UTC 5 Dec 2001 at a temperature of −19°C; (b) same as in (a), but at 0127 UTC at −11°C; (c) 2D-C images simulated from silhouettes of idealized cold-type snow crystals shown in Fig. 7; and (d) examples of images classified as irregular from Korolev et al. (2000). The vertical width of all strips is 800 μm.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

Fig. 7.
Fig. 7.

Idealized cold-type snow crystals from Magono and Lee (1966, their Fig. 1): (a) radiating assemblage of plates; (b) bullets with plates; (c) sideplanes; (d) scalelike sideplanes; and (e) combination of sideplanes, bullets, and columns.

Citation: Journal of the Atmospheric Sciences 64, 7; 10.1175/JAS3962.1

1

Although broken branches can usually be identified as dendritic, stellar, sectorlike, etc., it is not possible to distinguish the types of the parent crystal, in terms of monocrystalline versus polycrystalline or single-habit versus multihabit. Therefore, it is appropriate to retain such fragments in the irregular category. Alternatively, separate categories for different types of broken branches could be defined, but Magono and Lee (1966) did not do this, and we adhere to their convention.

Save