Open Access Atmos. Chem. Phys., 14, 1649–1662, 2014 Atmospheric www.atmos-chem-phys.net/14/1649/2014/ doi:10.5194/acp-14-1649-2014 Chemistry © Author(s) 2014. CC Attribution 3.0 License. and Physics

Incidence of rough and irregular atmospheric ice particles from Small Ice Detector 3 measurements

Z. Ulanowski1, P. H. Kaye1, E. Hirst1, R. S. Greenaway1, R. J. Cotton2, E. Hesse1, and C. T. Collier1 1Centre for Atmospheric and Instrumentation Research, University of Hertfordshire, Hatfield AL10 9AB, UK 2Met Office, FitzRoy Road, Exeter EX1 3PB, UK

Correspondence to: Z. Ulanowski ([email protected])

Received: 2 September 2013 – Published in Atmos. Chem. Phys. Discuss.: 25 September 2013 Revised: 17 December 2013 – Accepted: 8 January 2014 – Published: 12 February 2014

Abstract. The knowledge of properties of ice crystals such tions that climate could be modified through aerosol seeding as size, shape, concavity and roughness is critical in the con- to reduce cirrus cover and optical depth, as the seeding may text of radiative properties of ice and mixed-phase clouds. result in decreased shortwave reflectivity. Limitations of current cloud probes to measure these proper- ties can be circumvented by acquiring two-dimensional light- scattering patterns instead of particle images. Such patterns were obtained in situ for the first time using the Small Ice 1 Introduction Detector 3 (SID-3) probe during several flights in a variety of mid-latitude mixed-phase and cirrus clouds. The patterns Cloud feedbacks remain the largest source of uncertainty in are analysed using several measures of pattern texture, se- climate models. In particular, uncertainties exist concerning lected to reveal the magnitude of particle roughness or com- the radiative forcing of clouds containing ice crystals, most plexity. The retrieved roughness is compared to values ob- notably cirrus. Indeed, whether cirrus clouds warm or cool tained from a range of well-characterized test particles in the the Earth surface depends critically on ice crystal morphol- laboratory. It is found that typical in situ roughness corre- ogy, among other factors. Reducing this uncertainty requires sponds to that found in the rougher subset of the test particles, detailed in situ characterization of cloud particles, so that the and sometimes even extends beyond the most extreme values scattering properties of the clouds can be correctly repre- found in the laboratory. In this study we do not differentiate sented in models. Also, detailed knowledge of the scatter- between small-scale, fine surface roughness and large-scale ing properties of various cloud particle types is needed for crystal complexity. Instead, we argue that both can have sim- accurate retrieval of cloud microphysical properties from re- ilar manifestations in terms of light-scattering properties and mote sensing (Baran, 2012; Guignard et al., 2012). One of also similar causes. Overall, the in situ data are consistent, the main barriers to achieving these goals is the inability of with ice particles with highly irregular or rough surfaces be- cloud probes to determine the contribution of small ice crys- ing dominant. Similar magnitudes of roughness were found tals (that is crystals smaller than about 50 µm) to the total dis- in growth and sublimation zones of cirrus. The roughness tribution. This is due to crystal breakup on the inlets of these was found to be negatively correlated with the halo ratio, but probes (Field et al., 2006) and their inability to resolve pre- not with other thermodynamic or microphysical properties cisely the size and shape of small ice crystals because of the found in situ. Slightly higher roughness was observed in cir- conflicting demands of high optical resolution and large sam- rus forming in clean oceanic air masses than in a continental, ple volume (Ulanowski et al., 2004; Connolly et al., 2007; polluted one. Overall, the roughness and complexity are ex- Kaye et al., 2008; Bailey and Hallett, 2009). pected to lead to increased shortwave cloud reflectivity, in There is also growing evidence, most of it indirect or comparison with cirrus composed of more regular, smooth from remote sensing, that atmospheric ice crystals tend to ice crystal shapes. These findings put into question sugges- have shapes departing from idealized geometries based on perfect hexagonal prisms. Typically, aircraft and satellite

Published by Copernicus Publications on behalf of the European Geosciences Union. 1650 Z. Ulanowski et al.: Incidence of rough and irregular ice particles measurements of cirrus radiances show featureless phase Tracing with Diffraction on Facets (RTDF) model (Clarke functions, which are not consistent with the idealized geome- et al., 2006; Kaye et al., 2008; Hesse et al., 2009, 2012). Ice tries (Foot, 1988; Baran et al., 2001; Field et al., 2003; Baran, particle roughness can also be obtained, as evidenced by ex- 2004; Garrett, 2008; Baran, 2012). Korolev et al. (2000) con- perimental patterns from fluorosilicate ice analogue crystals cluded that the majority of ice particles in mid-latitude strat- with smooth and rough surfaces, which show distinct differ- iform clouds observed during several campaigns were of ir- ences: while the former have sharp, well-defined bright arcs regular shape. Roughness may have been responsible for un- and spots, the latter have much more random, “speckled” ap- expectedly low values of the asymmetry parameter measured pearance (Ulanowski et al., 2006). The average size of the in ice clouds by Garrett et al. (2001). Gayet et al. (2011) speckle spots is inversely proportional to particle size, offer- found prevalent particles with imperfect or complex shapes ing a calibration-free method for particle sizing (Ulanowski at the trailing edge of mid-latitude frontal cirrus, while the et al., 2012). leading edge was dominated by more regular plates that pro- Here we focus on the application of 2-D scattering pat- duced a 22◦ halo signature. For one full day of PARASOL terns to retrieving the roughness of atmospheric ice parti- data over ocean, very rough faceted particles provided an cles. We use the first airborne data from the SID-3 probe, ob- improved fit to polarized reflectances (Cole et al., 2013). In tained in mid-latitude cirrus and mixed-phase clouds during a study by Baum et al. (2011) some space-borne depolar- the UK Met Office CONSTRAIN campaign in 2010. Several ization lidar measurements could be explained by modelling features of the patterns are used to assess the roughness of ice crystals with rough surfaces. Best fits to data obtained by the corresponding ice particles. The features are compared to Lampert et al. (2009) in Arctic ice cloud were consistent with those obtained from natural and artificial particles of varying deeply rough hexagonal ice crystals. Measurements using the surface roughness and complexity. We then examine possi- polar nephelometer indicated that the surface of Antarctic ble causes of the observed roughness. In addition to growth ice crystals was deeply rough (Shcherbakov et al., 2006) . In processes we also consider sublimation, since ice surfaces Arctic mixed-phase clouds Jourdan et al. (2010) found a mix can become rough during rapid sublimation (Cross, 1969; of droplets and deeply rough droxtal-shaped ice crystals; but Kobayashi and Ohtake, 1974; Pfalzgraff et al., 2010). This is the droxtal particles were considered as surrogates for small contrary to what currently appears to be the prevailing view, irregular ice crystals. which postulates that atmospheric ice crystals merely be- It is important in this context that particle roughness can come rounded (Nelson, 1998). We investigate this dilemma dramatically alter the scattering properties of ice crystals. For by examining in situ cloud data as well as laboratory cases of example, it can significantly reduce the asymmetry parame- crystal sublimation and dissolution. ter (Yang et al., 2008; Ulanowski et al., 2006). Roughness Due to limitations of in situ cloud probes, little is known may also account for the relative rarity of ice halos (Yang about the detail of surface structure of atmospheric ice crys- and Liou, 1998; Mishchenko and Macke, 1999; Ulanowski, tals, although some attempts have been made to quantify it 2005). Therefore it is important to quantify fine detail of in the laboratory (Neshyba et al., 2013), and it has been pos- ice crystal geometry, currently beyond the reach of imaging tulated that it may be anisotropic in nature (Shcherbakov, cloud probes (Ulanowski et al., 2004; Connolly et al., 2007; 2013). Speckle techniques for the quantitative measurement Bailey and Hallett, 2009), although there are indications that of roughness depend on assumptions about the vertical sur- a new generation of probes using incoherent light may be face height distribution (roughness amplitude or variance) able to resolve at least some of the detail (Schön et al., 2011). and autocovariance function (horizontal correlation length). It is possible to circumvent the optical resolution limita- Moreover, particle surface roughness, as represented by pa- tions of imaging probes by acquiring light-scattering “pat- rameters such as the amplitude of variance of surface height, terns” instead of images. Such patterns can be obtained from is not in general related in a simple, linear manner to speckle relatively large sample volumes, as there is no sharply de- properties, except for low roughness, smaller than the wave- fined image plane to limit resolution. Several light-scattering length of light (Goodman, 2007). For these reasons, fully cloud probes, jointly known as Small Ice Detectors (SID) quantitative measurement of roughness cannot yet be ob- have been developed over the last decade at the Univer- tained with the available data. Consequently, in this work we sity of Hertfordshire. Successive models obtain scattering focus on semi-quantitative, exploratory analysis and compar- patterns with progressively higher angular resolution. The isons between data from atmospheric ice and test particles earlier designs rely on multi-element detectors measuring with known geometry. Furthermore, whilst the average area mainly the azimuthal scattering, while the most recent, of speckle spots is known to be inversely proportional to the collectively known as SID-3, acquire high-resolution two- size of the scattering particle, the width of the intensity “en- dimensional (2-D) scattering patterns (Kaye et al., 2008). velope” of the speckle pattern is expected to be related to Two-dimensional scattering patterns offer high potential for the transverse scale of surface roughness (Ulanowski et al., detailed particle characterization. It is possible to recover the 2012). This may in the future allow the recovery of such shape, size and orientation of small ice particles by com- scale from speckle patterns, but further research is needed paring such patterns to scattering models such as the Ray in this area. In the meantime, we limit ourselves to treating

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/ Z. Ulanowski et al.: Incidence of rough and irregular ice particles 1651

Fig. 1. 5-day backward air mass trajectories for cirrus flight B504 Fig. 2. As Fig. 1 but for cirrus flight B558 ending at 12:00 UTC on ending at 12:00 UTC on 27 January 2010, computed for final alti- 28 September 2010, computed for final altitudes of 8500, 9500 and tudes of 9000, 9500 and 10 000 m. The lower panel shows the alti- 10 500 m. tude above ground level in metres.

a congested, supercooled cumulus; it was profiled up to cloud in the same way particles that may be characterized by dif- top near 3 km. Most of the flights were in prevailing oceanic ferent transverse scales of roughness. This may include fine, air mass, with the exception of B558. Air mass origin was wavelength-scale roughness at one extreme, and large-scale determined using trajectory analysis. Five-day back trajecto- complexity at the other extreme. ries were computed using the NOAA HYSPLIT model with GDAS meteorological data, using vertical model velocities (Draxler and Rolph, 2003). The trajectories for flights B504 2 Methods (oceanic air mass) and B558 (continental air mass) are shown in Figs. 1 and 2, respectively. The first in situ cloud data from the SID-3 probe were ob- SID-3 was flown in a PMS-style canister on the FAAM tained during the UK Met Office CONSTRAIN campaign in research aircraft. The probe has “open” geometry similar to Scotland in January and February 2010, with another case, SID-2 – Fig. 3. This design minimizes the incidence of ice flight B558, in south-eastern England in September 2010. particles shattered on the probe housing (Cotton et al., 2010, Three flights in this analysis – B503 (25 January), B504 2013). Specifically, during CONSTRAIN shattering was not (27 January) and B558 (28 September) – sampled cirrus, thought to be important for SID-2 (Cotton et al., 2013). Since and two, B505 (29 January) and B509 (16 February) mixed- SID-3 has very similar geometry, it is unlikely that shattering phase clouds. The cloud sampled was mostly mid-latitude could have influenced the measurements significantly. Parti- stratiform cirrus associated with fronts within cyclonic sys- cle triggering (two photomultipliers defining the sample vol- tems. The aircraft sampling pattern consisted of a series of ume), incident illumination (532 nm wavelength laser beam) level runs each separated in altitude by 1000 ft, and ranged and sample volume definition are also similar to SID-2 (Cot- from near cloud top to below cloud base. The cirrus cloud ton et al., 2010). However, the main detector of SID-3 is an top temperatures were around −60 ◦C and the cloud base intensified CCD camera (Photek Ltd, UK) producing images around −30 ◦C. Observations near the cloud base included 780 by 582 pixels in size. The gain of the intensifier, and sublimating particles. The relative humidity with respect to hence the resulting image brightness, can be varied under ice, using the General Eastern chilled mirror hygrometer was software control. The camera delivers 2-D scattering patterns mostly between 80 and 110 % throughout the depth of the from single particles at rates up to 30 per second, depending cloud. The highest altitude reached during these flights was on configuration. Since the sampling volume for the airborne 10.9 km. Flight B509 was also in a frontal origin cloud, but instrument is typically about 50 mLs−1 in flight, not all par- its base was much lower (≈ 1 km) so it was mixed phase; it ticles can be captured as 2-D patterns at higher concentra- was profiled up to cloud top near 4.5 km. Flight B505 was in tions. However, data from the trigger channels are recorded, www.atmos-chem-phys.net/14/1649/2014/ Atmos. Chem. Phys., 14, 1649–1662, 2014 1652 Z. Ulanowski et al.: Incidence of rough and irregular ice particles

Images of ice crystals were obtained in the presence of wa- ter vapour in the SEM chamber (pressure ≈ 40 Pa) with only residual amounts of air. Ice was grown from water vapour on a metal substrate on a Deben Ultra Peltier-cooled cold stage, and sublimated by increasing the substrate temperature. Image texture can be quantified using statistical measures, e.g. the grey-level co-occurrence matrix (GLCM), which deals with spatial relationships of pairs of grey-value pix- els (Haralick et al., 1973). Previously, GLCM was applied to retrieving surface roughness from laser speckle images (Lu et al., 2006). Initially, four GLCM features were chosen: con- trast, correlation, energy and homogeneity. They were calcu- lated for nearest neighbour pixels in four directions (Haral- ick et al., 1973). In addition, image entropy and two mea- sures relating to image brightness distribution, rather than texture, were examined: the ratio of root-mean-square bright- Fig. 3. Schematic view of the head of the SID-3 instrument. Only ness (RMS) to its standard deviation (SD) (RMS / SD; Jolic a small part of the instrument inside the PMS canister is visible et al., 1994), and kurtosis (defined as fourth moment about (lower right). The sensing area shown is positioned 17.4 mm from the mean divided by the 4th power of the standard devia- the faceplate of the instrument. tion). We emphasize that, while the GLCM depends on spa- tial structure of the image (relative brightness of neighbour- ing pixels), RMS / SD and kurtosis do not (they describe only including amplitude histograms, allowing some reconstruc- the statistical distribution of brightness over the entirety of tion of particle number concentrations. Receiving optics col- the image), so they express somewhat different properties lect the scattered light over an annulus covering scattering of speckle. To calculate the measures, only the pixels from angles from 6 to 25◦, sufficient to encompass the 22◦ halo SID-3 image within the 6 to 25◦ scattering angle range were scattering from ice prisms, but with the central low-angle used, i.e. only within the image annulus. The measures were area obscured by a beam stop. Several instruments based on calculated for SID-3 2-D patterns obtained from a range of this principle have been constructed, including both airborne test particles: smooth and rough ice analogues and mineral and laboratory versions (Kaye et al., 2008). The camera im- dust grains, and correlated with a semi-quantitative measure ages of the 2-D patterns were digitized in this study as 8-bit of particle roughness. We also examined the sensitivity of compressed JPEG files, although the system is also capable the same measures to potential bias sources, including image of producing lossless 12-bit TIFF files. Since the 8-bit bright- noise, saturation, particle size, shape and orientation, and im- ness is quite narrow in terms of dynamic range, and the in- age intensifier gain (hence image brightness). Various image tention was to collect images from small as well as larger ice normalization and averaging schemes were compared too, particles, the intensifier gain was switched every three sec- and the chosen one involved scaling mean image brightness onds between two values differing typically by a factor of to 10 on the 0–255 scale. 10. Of the four GLCM features, energy was found to have the In addition to in situ data, SID-3 2-D patterns were also strongest correlation with particle roughness and was most obtained from a variety of single test particles in the labo- robust with respect to the potential bias sources. It is rele- ratory. For these measurements, rough and smooth mineral vant that the GLCM energy also shows good correlation with dust grains (Ulanowski et al., 2012) and ice analogue crys- roughness in the context of laser speckle from flat surfaces, tals (refractive index of 1.31) with smooth surfaces, rough and is most robust with respect to variation of “the setup con- surfaces or complex structure were used (Ulanowski et al., figuration, the position, and the orientation of the surface to 2003, 2006). They were deposited on anti-reflection coated be measured” (Lu et al., 2006). So only this GLCM feature is glass windows with specified reflectance < 0.2 % per surface. reported henceforth. Similar performance characterized the Test particle size was obtained from optical microscopy im- remaining measures: entropy, RMS / SD and kurtosis; but ages taken in the same particle orientation as in the scatter- different measures showed sensitivity to different sources of ing data. Most of these particles were also visualized using bias. For example, entropy was more sensitive to gain, par- scanning electron microscopy (SEM). Details of the particles ticle size and noise in the images, so it was not included are given in Table 3. SEM images of most of the test par- in further analysis. Kurtosis on the other hand was sensi- ticles were taken using a JEOL-5700 environmental SEM, tive to detector gain (image brightness). A “combined rough- and of dissolved ice analogues using a Camscan CS44 SEM. ness” measure was also defined, composed of the most robust All SEM images were obtained with backscattered electron measures, energy E, RMS / SD and the base 10 logarithm of detectors under low vacuum and without specimen coating.

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/ Z. Ulanowski et al.: Incidence of rough and irregular ice particles 1653

Fig. 4. Combined roughness measure of test particles plotted against subjective roughness (see Sect. 2). A wider range of particles than in Table 3 was tested. Fig. 5. Six randomly selected SID-3 patterns from ice particles seen during CONSTRAIN cirrus (top) and mixed-phase (bottom) flights. kurtosis K, as follows: 0.7 − 2E/3 − (logK)/6 + (RMS/SD)/4000. (1) hygrometer, built and maintained by the UK Met Office (Keramitsoglou et al., 2002; Fahey et al., 2009). The signs in front of the individual terms reflect the signs The so-called “halo ratio” (HR) is a quantitative criterion of the correlations with roughness: negative for energy and to characterize the occurrence of the 22◦ halo peak. It is de- kurtosis and positive for RMS / SD. fined by the ratio of the scattered energy values measured The combined roughness measure was weighted so that at the scattering angles of 22◦ and 18.5◦. As such, it is sen- the individual measures contributed to it approximately sitive to the presence in ice crystals of the 60◦ prism angle equally and it had bounds of 0 and 1 for most particles en- and to the smoothness of the prismatic faces (Auriol et al., countered (although these bounds can be exceeded for very 2001; Gayet et al., 2011). Here it was calculated by averaging smooth or very rough particles). The combined roughness SID-3 image brightness between the scattering angles from was characterized by reduced sensitivity to image brightness, 21.5◦ to 22.5◦, and from 18◦ to 19◦, and taking the ratio of although it still showed sensitivity to image saturation. To re- the respective averages. To aid azimuthal averaging, the im- duce the impact of the residual bias sensitivities, very dark ages were first converted from Cartesian to polar coordinates, or excessively saturated patterns were removed from further using triangle-based linear interpolation. analysis: minimum brightness was set at 4 and maximum number of saturated pixels at 1000 (excluding the central beam stop). 3 Results The selected image features were calculated for 2-D pat- terns from cirrus and mixed-phase cloud particles as well as Figure 5 shows a typical selection of SID-3 scattering pat- test particles including mineral dust grains and smooth and terns from ice particles from the CONSTRAIN flight cam- rough ice analogues. The measures obtained from the test paign. The roughness measures were calculated for a random particles were correlated with a subjective measure of par- selection of several hundred patterns from marine cirrus (239 ticle roughness derived from microscopy images – Fig. 4 patterns) and mixed-phase (199 patterns) flights, as well as shows the correlation with the combined roughness. We also one cirrus flight in a continental air mass (59 patterns). The attempted to retrieve surface profiles of the test particles from roughness measures are summarized in Table 1 and the fre- stereo images taken using SEM, but these proved to be inac- quency distributions are shown in Fig. 6. Statistical signifi- curate, especially for the somewhat heterogeneous mineral cance tests for differences between the roughness measures dust grains, unless destructive specimen preparation proce- were carried out using the t test method, assuming unequal dures were used, so this route was abandoned in order to variances. The probabilities that the null hypothesis of no dif- maintain intact test particle collection for further work. Ice ference between the measures is true are given in Table 2. particle size was retrieved from the SID-3 patterns by de- The roughness measures were also calculated for a range termining average speckle area, which is inversely propor- of test particles, including ice analogue crystals (Ulanowski tional to particle size, using a method described previously et al., 2006) and mineral dust grains, all representing a range (Ulanowski et al., 2012). of surface roughness – the results are given in Table 3, and the Relative humidity with respect to ice was determined from magnitudes and range of the roughness measure values of the water vapour concentration measured on FAAM using the test particles are shown graphically above the frequency dis- Fluorescent Water Vapor Sensor (FWVS) fast Lyman-alpha tributions in Fig. 6. Images of the test particles from Table 3, www.atmos-chem-phys.net/14/1649/2014/ Atmos. Chem. Phys., 14, 1649–1662, 2014 1654 Z. Ulanowski et al.: Incidence of rough and irregular ice particles

Table 1. Mean values and standard deviations (in brackets) of roughness measures of ice crystals in the three cloud cases shown in Fig. 6, obtained from 2-D SID-3 scattering patterns.

Case Energy logK RMS / SD Combined roughness Marine cirrus 0.239 (0.084) 1.38 (0.39) 1100 (336) 0.586 (0.175) Continental 0.218 (0.062) 1.33 (0.25) 605 (139) 0.485 (0.103) Mixed phase 0.224 (0.0714) 1.20 (0.26) 936 (228) 0.586 (0.123)

Fig. 6. Frequency distributions of roughness measures obtained from SID-3 2-D patterns in marine air mass cirrus (top row), continental air mass cirrus (middle row), and mixed-phase clouds (bottom row) flights. Labelled lines above the distributions show roughness measure values for six test particles from Table 3 and Fig. 7: (a), (b), (d), (e), (g) and (j), where (a) is the smoothest and (j) the roughest one. Note that the GLCM energy and log kurtosis are anticorrelated with roughness. together with corresponding SID-3 scattering patterns, are to the presence of bright spots in the 2-D pattern at scattering shown in Fig. 7. A total of 23 test patterns from 16 particles angles > 22◦. were used in the comparison given here, as shown in Fig. 4. The roughness measures were examined for possible cor- Data from some of the particles were obtained for two ori- relation with other variables measured during the flights, in- entations and up to three different image intensifier gain val- cluding temperature, ice supersaturation, vertical velocity, ues. For example, for particle (b) and (f) there were two gains ice particle size, ice particle number concentration and large and two orientations, for particle (e) two orientations, and for aerosol concentration, although not all of these variables particle (f) three gains and two orientations. The variability were available simultaneously for all flights. Generally, the of roughness measures can be seen in Fig. 4 by looking at the correlations were weak – an example is given in Fig. 8, which spread of data points along the vertical axis. Different orien- shows roughness plotted against relative humidity with re- tations were included because we had found that the speckle spect to ice. In contrast, stronger negative correlation was measures varied with particle orientation. For example, for found between roughness and the halo ratio determined from the particles (b), (e) and (f) the combined roughness varied the SID-3 patterns – see Fig. 9. The least-squares regres- with orientation between 0.32 and 0.38, 0.34 and 0.7, and sion between the halo ratio (HR) and RMS / SD was HR = 0.78 and 0.87, respectively. We interpret this variability as 1.11 − 0.125 × 10−3 RMS / SD. originating partly or mostly from the fact that the roughness We also carried out initial investigations of possible pro- measures depend on the presence or absence of features in cesses that may lead to the formation of rough ice surfaces, the 2-D scattering patterns: for smooth crystals bright spots using ice as well as ice analogues as materials. a scanning and arcs appear (Ulanowski et al., 2006), and their presence electron microscopy image of ice crystals grown and then leads to lowered roughness measures. However, the position partially sublimated in the presence of water vapour in the of these features depends on crystal orientation, and in some microscope chamber is shown in Fig. 10. It is noticeable orientations they may disappear entirely from the 6 to 25◦ that surfaces deeper within the sample remained smooth, scattering angle range encompassed by the SID-3 detector. while the exposed ones become roughened, and that the pris- For instance, the existence of the 22◦ halo peak is equivalent matic facets showed grooves perpendicular to the crystal c axis, while the roughness on the basal facets was isotropic.

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/ Z. Ulanowski et al.: Incidence of rough and irregular ice particles 1655

Fig. 7. Upper rows: images of test particles from optical microscopy (ice analogue columns (a) and (d) only) or SEM, labelled with the ID letters from Table 3, with size scales shown in insets – note that the magnifications vary; the dust grain (i) is shown in two magnifications. Lower rows: corresponding SID-3 scattering patterns.

Table 2. Significance tests for differences between the roughness measures of ice crystals in the three cloud cases shown in Fig. 6 (marine cirrus, continental cirrus, mixed phase), obtained using the t test method. Shown are the probabilities that the null hypothesis of no difference between the measures is true (low probabilities correspond to more significant differences). Values < 5 % are given in bold to highlight significant differences.

Energy logK RMS / SD Combined roughness Case A Case B (%) (%) (%) (%) Marine Continental 3.9 18 4.3 × 10−42 5.0 × 10−6 Marine Mixed phase 4.2 4.7 × 10−7 3.0 × 10−7 99 Continental Mixed phase 59 6.3 × 10−2 1.4 × 10−37 6.5 × 10−7

Overall, we found that, while irregular growth could be pro- the resolution, was higher than for imaging cloud probes, so duced under high supersaturation, in general fine roughness the latter would reveal even less detail. was easier to obtain by sublimation than by growth. Typi- cal SEM images of ice analogues grown from solution under low and high supersaturation are represented by the test par- 4 Discussion ticles “a” or “b”, and “g”, respectively, in Fig. 7. The former pair shows regular hexagonal shapes with smooth, flat facets, We have presented here the first, exploratory results from in while the latter crystal is highly deformed. Regular ice ana- situ measurements using the SID-3 cloud probe, which al- logue crystals were partially dissolved by immersing them in lows ice particle characterization through the analysis of 2-D subsaturated sodium fluorosilicate solution, while observing scattering patterns, instead of images. We have found that the them continuously using optical microscopy. A typical ex- scattering patterns obtained in both cirrus and mixed-phase ample is shown in Fig. 11: following dissolution at 99 % sat- clouds during the CONSTRAIN campaign in 2010 were typ- uration the crystal could be seen to attain oval outline, with ically characterized by the presence of pronounced speckle. only a hint of underlying fine structure. However, the higher- Qualitatively, these patterns are similar in appearance to pat- resolution SEM image of another crystal from the same batch terns obtained from the rougher particles in the test particle shown in Fig. 12 reveals much fine detail, including pro- selection shown in Fig. 7. Comparison of the distributions of nounced ridges and cavities. It must be noted that the nu- in situ roughness measures shown in Fig. 6 with roughness merical aperture of the optical microscope (0.4), and hence values for the test particles reveals that smoother particles corresponded to the tails of the roughness distributions. The www.atmos-chem-phys.net/14/1649/2014/ Atmos. Chem. Phys., 14, 1649–1662, 2014 1656 Z. Ulanowski et al.: Incidence of rough and irregular ice particles

Table 3. Ice analogue and mineral dust test particles in increasing order of subjective roughness (see Sect. 2). The size is the approximate maximum dimension. The roughness measures are GLCM energy E, log kurtosis logK, RMS / SD and “combined roughness” (see text); the signs in brackets indicate the sign of correlation with roughness.

ID Type Size E logK RMS / SD Combined Subjective (µm) (−)(−)(+)(+) roughness a Analogue column 140 0.43 2.18 683 0.22 1 b Analogue rosette 150 0.34 2.08 993 0.37 2 c Dust, smooth 65 0.32 2.15 733 0.31 4 d Analogue column 160 0.31 1.60 1059 0.49 4 e Analogue rosette 51 0.16 1.03 1107 0.70 5 f Analogue rosette 160 0.29 1.42 1171 0.56 5 g Analogue plate 30 0.22 1.08 1052 0.64 6 h Dust, rough 35 0.23 1.26 1294 0.68 8 i Dust, rough 47 0.24 1.12 1088 0.62 9 j Dust, rough 46 0.20 1.07 1298 0.71 10 distributions were better represented by the rougher or more ticles. The 2-D patterns obtained for test particles show that irregular particles in the selection, such as the highly dis- substantial speckle can result either from small-scale, fine torted ice analogue prisms (d and g), the “germ” or rough ice roughness or from complexity – the former example being analogue rosettes (e and f) or the rough mineral dust grains mineral dust grains, the latter the complex, “germ rosette” (h, i and j). The “rough” tails of the roughness measure distri- with smooth prismatic facets. This similarity is not surpris- butions extended in some cases beyond the roughness values ing, as speckle will arise if multiple “scattering centres” – obtained for the roughest particles tested. mutually incoherent radiating regions – are present on the Phase functions of smooth and rough ice analogue rosettes surface of the particle (Goodman, 2007; Ulanowski et al., very similar to those used here was previously measured us- 2012). This can occur if particle surfaces are rough, or the ing a levitation technique. It was found that the transition geometry of the particle is very complex, so that scattering from smooth to rough geometry for these large (≈ 90 µm in any given direction contains contributions from different maximum dimension) crystals lowered the asymmetry pa- particle regions, leading to interference. The same should ap- rameter from 0.81 to about 0.63 (Ulanowski et al., 2006). ply to hollow crystals, as additional facets would be present. It is worth noting that such a large change corresponds to al- Moreover, for speckle to occur, the wavefront delays (phase most doubling the reflectivity of a cloud composed entirely of differences) introduced into the scattered wave by these re- such particles. Similar lowering of the asymmetry parameter gions must be comparable to or greater than the wavelength between smooth and rough particles was found for mineral of light; but what matters is the difference between the de- dust grains (McCall, 2011). lay and the whole number of wavelengths, rather than the The halo ratio calculated from scattering phase functions absolute delays (the phase is “wrapped” modulo 2π). There- is a measure of the quality of the 22◦ halo peak and as such fore, both small and large departures from surface flatness indicates ice particle “regularity”. The halo ratio and asym- can influence scattering in similar ways. a corollary of the metry parameter measured in situ are positively correlated presence of such phase-delaying centres in the case of ice (Auriol et al., 2001; Gayet et al., 2011). The halo ratios of the crystals is that parallel facets are no longer present. Parallel sets of smooth and rough ice analogue crystals measured pre- facets give rise to unscattered “delta rays” (δ-function trans- viously (Ulanowski et al., 2006) were 1.40 and 0.89, respec- mission) in the large particle (geometric optics) limit (Takano tively. We also found negative correlation between roughness and Liou, 1989; Mishchenko and Macke, 1998) or forward- and the halo ratio determined in cirrus (Fig. 9). The correla- scattered light in general (once diffraction has been properly tion is statistically highly significant. Moreover, the halo ratio accounted for). So the absence of parallel facets leads to in- was determined for single particles, each in its own orienta- creased scattering at larger angles and, in turn, lower asym- tion; as such, the ratio is subject to very wide variation due metry parameter, due to its cosine weighting factor (Nousi- to the orientation, which is likely to cause most of the ob- ainen et al., 2011). served variability. Had the halo ratio been determined for an Thus overall, roughness and complexity appear to im- ensemble of particles (as is the case for the polar nephelome- pact not just 2-D scattering but other scattering proper- ter), the variability is likely to have been much lower. The ties in a similar manner. The reduction in the asymme- presence of the correlation reinforces the expectation that try parameter is shown by both measurement and mod- negative correlation should be present between the retrieved elling studies (Ulanowski et al., 2006; Yang et al., 2008; roughness and the asymmetry parameter of the in situ ice par- Um and McFarquhar, 2011). In a manner, whatever the

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/ Z. Ulanowski et al.: Incidence of rough and irregular ice particles 1657

Fig. 9. Halo ratio plotted against particle roughness measure Fig. 8. Relative humidity with respect to ice measured using the RMD/SD, both determined from SID-3 patterns obtained in cirrus FWVS hygrometer on the FAAM aircraft during the cirrus flight formed in marine air masses. Mean halo ratio calculated for fixed B504, plotted against ice particle roughness as represented by the roughness intervals, and a least-squares regression line (coefficients “combined roughness” measure (points) and mean roughness (line). given in the legend) are also shown. The correlation coefficient was ◦ Average temperature during the relevant flight sections was −54 C. −0.22, significant at the 0.1 % level, based on 239 data points.

“roughness” structure, large rough particles acquire some non-spherical shapes like ellipsoids generally have dramat- scattering properties that are more characteristic of small par- ically different scattering properties from complex, facetted ticles, i.e. strong side-scattering and low asymmetry parame- particles, including larger asymmetry parameter (Ulanowski ter (Ulanowski et al., 2006; Um and McFarquhar, 2011). et al., 2004; Connolly et al., 2007; Um and McFarquhar, Concerning the mineral dust grains as surrogates for ice 2011). Thus it is very important to correctly identify such particles, while the refractive index of mineral dust is higher, particles from in situ data and to represent their scattering the difference in roughness that would be caused by such properties appropriately in models. Speckle analysis offers change is small: if dust is substituted for ice, the same speckle the first step in this process, and while discrimination from pattern should be obtained if dust roughness is 1.31/nm of particles with fine roughness is not yet carried out, it may that of ice (Ruiz Gale et al., 2007), where nm, the refractive not be essential due to the similarity of the scattering prop- index of mineral dust, is typically ≈ 1.52. So the dust parti- erties. Moreover, as the above discussion illustrates, small- cles may represent both the scattering properties and possibly scale roughness and large-scale complexity can have similar also the shape of atmospheric ice crystal quite well, at least causes, for example the presence of high water vapour super- as far as solid (as opposed to hollow), simple (as opposed saturation with respect to ice, lessening further the need for to branched or aggregated) crystals are concerned. But we detailed categorization. must keep in mind the caveats concerning the possible am- Possible causes of the prevalent ice crystal roughness biguity between rough and complex particles. At the same may be related to details of the ice crystal growth process, time we note that the roughness magnitudes found in situ ex- the level of water vapour supersaturation, ice sublimation, ceeded those found for the roughest test particles we could and also the type of nucleation. Homogeneously frozen ice procure; this suggests that both large-scale complexity and can contain high concentration of stacking faults, leading to high small-scale surface roughness may have been present partly ordered crystalline structure, neither hexagonal nor cu- in at least some ice particles. An example of such a particle bic. Such structure can be expected to lead to poorly defined might be a rosette or aggregate with very rough surfaces – facet geometry and the emergence of rough surfaces (Malkin the type of particle not present in our test collection. et al., 2012). Imaging cloud particle probes have difficulty resolving not Nelson (1998) suggests that repeated cycles of supersatu- only fine surface roughness but also larger detail of complex ration and undersaturation are likely to lead to more complex, particles. Complex near-spherical crystals, which could po- “exaggerated” crystal shapes, as a consequence of asym- tentially take on a variety of possible shapes – such as, for ex- metry between growth and sublimation. Therefore mixing ample, germ rosettes (“budding Bucky balls”) – are typically and small-scale vertical motions in clouds, leading to the classified simply as “irregulars”, and in the past were even presence of many growth and sublimation regions, may be categorized as spheres. However, spheres or even compact, a possible mechanism by which roughness emerges (Korolev www.atmos-chem-phys.net/14/1649/2014/ Atmos. Chem. Phys., 14, 1649–1662, 2014 1658 Z. Ulanowski et al.: Incidence of rough and irregular ice particles

Fig. 11. Optical microscopy images of an ice analogue crystal in subsaturated sodium fluorosilicate solution at the start of dissolv- ing (left) and after approximately 20 min (right). Initial crystal di- mensions were 120 by 37 µm. Phase contrast was used and contrast stretching applied to emphasize the small difference in the refractive index of the solution (1.33) and the crystal (1.31). The numerical aperture of the objective was 0.4.

Fig. 10. SEM image of ice crystals sublimating at a temperature of −30 ◦C.

et al., 1999). Indeed, high-resolution modelling predicts that ice particle trajectories in mixed-phase clouds contain mul- tiple super- and subsaturated regions (Flossmann and Wo- Fig. 12. SEM images of an ice analogue crystal from the same batch brock, 2010). Thus one of the causes of the ice particle as in Fig. 11, before (left) and after approximately 11 min of dis- roughness observed in situ may be the cyclic nature of the solving (right). Initial crystal diameter was 50 µm. growth process, at least in the case of mixed-phase clouds. Furthermore, the same modelling shows that a broad range of humidity values can be present, departing widely from equi- shielded and, hence, were subjected to less severe subsatura- librium. In situ cirrus measurements too show a wide varia- tion because of lower temperature and/or higher local water tion of humidity, up to the homogeneous freezing threshold vapour pressure (Fig. 10). This suggests that roughness may and sometimes slightly above (Krämer et al., 2009). Prelim- emerge during sublimation only if ice crystals fall through inary results from experiments in a flow diffusion chamber, relatively dry air. On the other hand, if stacking faults were where ice crystals were grown on glass fibres and observed present, localized sublimation might even occur at vapour using a laboratory version of SID-3, also indicate increase pressures corresponding to supersaturation with respect to in roughness following repeated growth cycles (Voigtländer hexagonal ice, as cubic ice has higher vapour pressure than et al., 2013). hexagonal ice (Shilling et al., 2006). Roughening during sub- We observed substantial roughness on ice particles sub- limation could explain the apparent roughness of ice particles limating on a cold stage of an SEM (Fig. 10), in agree- we found in the sublimation zone of cirrus, even at low hu- ment with other work (Cross, 1969; Pfalzgraff et al., 2010; midities (Fig. 8). However, an alternative explanation is that Neshyba et al., 2013). The emergence of grooves on subli- the observed apparent roughness is in part due to increased mating prismatic faces may be due to the presence of stack- ice particle complexity caused by high aggregation rates in ing faults (Kobayashi and Ohtake, 1974; Kuhs et al., 2012). the sublimation zone (Sölch and Kärcher, 2011). Roughening was also seen when ice analogue crystals were The lack of correlation between ice supersaturation and dissolved rapidly (Fig. 12), with roughness superimposed on roughness, found during the cirrus flights in this study, may the overall ellipsoidal shape. These findings are in contrast to be because the presence of roughness is connected to the the prevailing view that sublimation leads merely to crystal history of supersaturation stretching over the whole crystal rounding (Nelson, 1998). Despite the somewhat larger scale growth period, rather than the instantaneous value during the of ice analogue roughening in comparison to the real ice sam- measurement. Similarly, no correlation is reported between ples, the analogues appeared more or less rounded in lower- the instantaneous supersaturation and the halo ratio (an indi- resolution optical microscopy images, adding to the evidence cation of crystal “imperfection”) determined from the polar that in situ cloud particle observations using imaging probes nephelometer measurements (Gayet et al., 2011). (which are characterized by even lower resolution) do not re- It is interesting to observe that, while rough particles dom- veal relevant surface detail due to resolution limitations. It is inated in all three cloud types, and marine cirrus and mixed- also significant that smooth ice surfaces were visible on SEM phase clouds were quite similar, cirrus in a continental, pol- images of the same samples where the crystals were partially luted airflow showed significantly lower roughness for all

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/ Z. Ulanowski et al.: Incidence of rough and irregular ice particles 1659 measures apart from kurtosis. We speculate that this was brighter if strong roughness was included, accordingly to the due to higher concentration of inhomogeneous ice nuclei aerosol hypothesis. However, we note that other factors may (IN) in the last case (unfortunately, comparative aerosol mea- also contribute to the discrepancy. surements were not available for these cases during CON- It has been proposed that climate could be modified STRAIN). through aerosol seeding to reduce cirrus cover and optical It is relevant in this context that high-resolution modelling depth (Mitchell et al., 2011). That such a strategy may be indicates higher in-cloud ice supersaturation at low IN con- ineffective or even counterproductive is shown by the ap- centrations in mixed-phase (Flossmann and Wobrock, 2010) parent decrease in roughness with IN concentration and the and cirrus clouds (Spichtinger and Gierens, 2009). Higher consequent increase in the asymmetry parameter seen in the supersaturations would lead to faster crystal growth, which present study, as it would result in decreased shortwave re- may in turn lead to higher incidence of structural defects, and flectivity of the modified cirrus. hence increased roughness. Low supersaturations are known to favour regular crystal growth, with solid, hexagonal prisms at the lower end of the superasaturation scale; conversely, 5 Conclusions shapes such as “polycrystals” or highly complex dendrites are common at the upper one (Peterson et al., 2010; Bai- The SID-3 probe was flown on the FAAM BaE-146 air- ley and Hallett, 2009, 2004). Incidentally, we found that this craft in mid-latitude clouds. Unlike most earlier results from was also true of the ice analogues: slow growth, for example cloud chambers, where SID-3 2-D scattering patterns typ- due to very slow evaporation of the saturated growth solu- ically displayed characteristics attributable to idealized ge- tion, produced regular, solid, prismatic crystals, while rapid ometric crystal shapes, the majority of the cloud patterns evaporation led to dendrites. On the other hand, lower IN showed random, “speckled” appearance. Lab experiments concentrations can also increase the contribution from ho- show that such appearance is typical of particles with rough mogeneous nucleation (e.g. Spichtinger and Gierens, 2009), surfaces or complex structure. Quantitative comparison of which could potentially influence crystal morphology, as the lab and cloud data was done using pattern texture measures, starting point for growth would be a frozen droplet, so the originally developed for surface roughness analysis using resulting crystal could become more complex – perhaps of laser speckle. The results are consistent with the presence of the irregular or the germ rosette type. Such crystals would strong roughness in the majority of cirrus and mixed-phase also produce strong speckle in the 2-D scattering patterns. cloud ice crystals, at levels similar to those found in rough Whatever the cause, we can term this conjecture the “aerosol ice analogue and mineral dust particles used for reference, hypothesis of roughness”. but some in situ roughness values were higher than for the Cloud chamber experiments focused on inhomogeneous “roughest” test particles we could obtain. Similar roughness nucleation tend to produce ice crystals with 2-D scatter- was found in the growth and sublimation zones of cirrus, ing features indicative of regular geometry and smooth sur- suggesting that roughness was maintained or possibly even faces (Ulanowski et al., 2007; Kaye et al., 2008), probably reinforced by sublimation. Slightly weaker roughness was as a result of relatively low supersaturations present dur- present in cirrus in a polluted air mass of continental ori- ing such experiments. However, there is accumulating evi- gin than in marine cirrus, possibly as an indirect outcome of dence from more recent, dedicated experiments in the AIDA higher IN concentrations in the former – “the aerosol hypoth- cloud chamber that increasing the maximum supersatura- esis of roughness”. The roughness was anticorrelated with tion achieved during chamber expansions leads to increased the halo ratio determined for single particles. roughness, and that homogeneous nucleation leads to crys- Finally, in terms of purely technical issues, our analysis tals with strongly rough surfaces, as indicated by SID-3 mea- was made difficult by the use of lossy JPEG images with surements (M. Schnaiter, personal communication, 2013). very limited, 8-bit dynamic range. This necessitated the use It is conceivable, therefore, that the excess of rough crys- of variable image intensifier gain to allow capturing data tals observed in the clean air mass cirrus flights was due from small as well as larger particles. These facts compli- to higher occurrence of homogeneous nucleation. The in- cated the data analysis and interpretation, and required the creased roughness could be a consequence of high incidence rejection of a significant proportion of the data, since all the of stacking faults in ice crystals, or increased complexity, as measures of roughness we used were found to be sensitive already discussed. to a greater or lesser extent to image brightness, noise, com- The lowering of the shortwave asymmetry parameter re- pression artefacts and saturation. SID-3 data from subsequent quired by the inclusion of strong roughness may improve the flying campaigns have been collected as lossless 12-bit TIFF agreement between global model predictions of cloud bright- files, without gain switching. These campaigns also provide ness. For example, the brightness predicted by some mod- more accurate and extensive supplementary measurements, els for the Southern Ocean region, where cirrus incidence such as water vapour concentration for deriving supersatura- is high, is too low (Baran, 2012). Since the air mass over the tion, and also aerosol measurement. Among others, this will Southern Ocean is relatively clean, this region would become allow more detailed study of the relationships between ice www.atmos-chem-phys.net/14/1649/2014/ Atmos. Chem. Phys., 14, 1649–1662, 2014 1660 Z. Ulanowski et al.: Incidence of rough and irregular ice particles roughness and humidity or IN concentration than was possi- Connolly, P. J., Flynn, M. J., Ulanowski, Z., Choularton, T. W., Gal- ble in this first, exploratory investigation. The analysis of the lagher, M. W., and Bower, K. N.: Calibration of 2-D imaging new data is already ongoing and will be reported in the near probes using calibration beads and ice crystal analogues, J. At- future. We have also initiated experimental and theoretical mos. Ocean. Tech., 24, 1860–1879, 2007. studies to determine the single-scattering properties, includ- Cotton, R., Osborne, S., Ulanowski, Z., Hirst, E., Kaye, P. H., and ing the asymmetry parameter, of particles with geometries Greenaway, R.: The ability of the Small Ice Detector, SID2 to characterise cloud particle and aerosol morphologies obtained corresponding to those that dominated the in situ measure- during flights of the FAAM BAe-146 research aircraft, J. Atmos. ments in the present study, so that the impact of our findings Ocean. Tech., 27, 290–303, 2010. on the radiative properties of cirrus and mixed-phase clouds Cotton, R. J., Field, P. R., Ulanowski, Z., Kaye, P. H., Hirst, E., can be quantified. Greenaway, R. S., Crawford, I., Crosier, J., and Dorsey, J.: The effective density of small ice particles obtained from in situ air- craft observations of mid-latitude cirrus, Q. J. Royal Meteor. Acknowledgements. This work was supported by the UK Natural Soc., 139, 1923–1934, 2013. Environment Research Council grants NE/E011225/1 (AP- Cross, J. D.: Scanning electron microscopy of evaporating ice, Sci- PRAISE) and NE/I020067/1 (ACID-PRUF). We are indebted ence, 164, 174–175, 1969. to Robert Williamson for skilful assistance in carrying out the Draxler, R. R. and Rolph, G. D.: HYSPLIT (HYbrid Single- ice analogue dissolving experiments and to Ken Henman for Particle Lagrangian Integrated Trajectory) model access via the electron microscopy of ice. We are grateful to the Facility NOAA ARL READY Website, available at: http://www.arl.noaa. for Airborne Atmospheric Measurements (FAAM) team for gov/HYSPLIT.php (last access: 29 August 2013), NOAA Air Re- making possible the measurements reported here and for providing sources Laboratory, Silver Spring, 2003. excellent support throughout the CONSTRAIN campaign. We also Fahey, D. W., Gao, R. S., and Möhler, O.: Summary of the AquaVIT gratefully acknowledge the NOAA Air Resources Laboratory for Water Vapor Intercomparison: Static Experiments, AquaVIT the provision of the HYSPLIT transport model. White Paper, Institute for Meteorology and Climate Research, Karlsruhe, available at: https://aquavit.icg.kfa-juelich.de/ Edited by: T. Garrett WhitePaper/AquaVITWhitePaper_Final_23Oct2009_6MB.pdf (last access: 22 August 2013), 2009. Field, P. R., Baran, A. J., Kaye, P. H., Hirst, E., and Greenaway, R.: References a test of cirrus ice crystal scattering phase functions, Geophys. Res. Lett., 30, 1752, doi:10.1029/2003GL017482, 2003. Auriol, F., Gayet, J.-F., Febvre, G., Jourdan, O., Labonnote, L., and Field, P. R., Heymsfield, A. J., and Bansemer, A.: Shattering and Brogniez, G.: In situ observations of cirrus cloud scattering phase particle interarrival times measured by optical array probes in ice ◦ ◦ function with 22 and 46 halos: cloud field study on 19 Febru- clouds, J. Atmos. Ocean. Tech., 23, 1357–1371, 2006. ary 1998, J. Atmos. Sci., 58, 3376–3390, 2001. Flossmann, A. I. and Wobrock, W.: A review of our understand- Bailey, M. and Hallett, J.: Growth rates and habits of ice crystals ing of the aerosol-cloud interaction from the perspective of a bin ◦ between −20 and −70 C, J. Atmos. Sci., 61, 514–544, 2004. resolved cloud scale modeling, Atmos. Res., 97, 478–497, 2010. Bailey, M. P. and Hallett, J. A.: Comprehensive habit diagram for Foot, J. S.: Some observations of the optical properties of clouds. atmospheric ice crystals: confirmation from the laboratory, AIRS II: Cirrus, Q. J. Roy. Meteor. Soc., 114, 145–164, 1988. II, and other field studies, J. Atmos. Sci., 66, 2888–2899, 2009. Garrett, T. J.: Observational quantification of the optical properties Baran, A. J.: On the scattering and absorption properties of cirrus of cirrus cloud, in: Light Scattering Reviews 3, Springer, Berlin, cloud, J. Quant. Spectrosc. Ra., 89, 17–36, 2004. 3–26, 2008. Baran, A. J.: From the single-scattering properties of ice crystals Garrett, T. J., Hobbs, P. V., and Gerber, H.: Shortwave, single- to climate prediction: a way forward, Atmos. Res., 112, 45–69, scattering properties of arctic ice clouds, J. Geophys. Res., 106, 2012. 15155–15172, doi:10.1029/2000JD900195, 2001. Baran, A. J., Francis, P. N., Labonnote, L.-C., and Doutriaux- Gayet, J.-F., Mioche, G., Shcherbakov, V., Gourbeyre, C., Boucher, M.: A scattering phase function for ice cloud: tests Busen, R., and Minikin, A.: Optical properties of pristine of applicability using aircraft and satellite multi-angle multi- ice crystals in mid-latitude cirrus clouds: a case study during wavelength radiance measurements of cirrus, Q. J. Roy. Meteor. CIRCLE-2 experiment, Atmos. Chem. Phys., 11, 2537–2544, Soc., 127, 2395–2416, 2001. doi:10.5194/acp-11-2537-2011, 2011. Baum, B. A., Yang, P., Heymsfield, A. J., Schmitt, C. G., Xie, Y., Goodman, J. W.: Speckle Phenomena in Optics: Theory and Appli- Bansemer, A., Hu, Y.-X., and Zhang, Z.: Improvements in short- cations, Roberts, Englewood, 2007. wave bulk scattering and absorption models for the remote sens- Guignard, A., Stubenrauch, C. J., Baran, A. J., and Armante, R.: ing of ice clouds, J. Appl. Meterol. Clim., 50, 1037–1056, 2011. Bulk microphysical properties of semi-transparent cirrus from Clarke, A. J. M., Hesse, E., Ulanowski, Z., and Kaye, P. H.: A 3- AIRS: a six year global climatology and statistical analy- D implementation of ray tracing combined with diffraction on sis in synergy with geometrical profiling data from CloudSat- facets, J. Quant. Spectrosc. Ra., 100, 103–114, 2006. CALIPSO, Atmos. Chem. Phys., 12, 503–525, doi:10.5194/acp- Cole, B. H., Yang, P., Baum, B. A., Riedi, J., Labonnote, L. C., 12-503-2012, 2012. Thieuleux, F., and Platnick, S.: Comparison of PARASOL obser- vations with polarized reflectances simulated using different ice habit mixtures, J. Appl. Meterol. Clim., 52, 186–196, 2013.

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/ Z. Ulanowski et al.: Incidence of rough and irregular ice particles 1661

Haralick, R. M., Shanmugam, K., and Dinstein, I.: Texture features Mishchenko, M. I. and Macke, A.: How big should hexagonal ice for image classification, IEEE T. Syst. Man Cyb., 3, 610–621, crystals be to produce halos?, Appl. Optics, 38, 1626–1629, 1973. 1999. Hesse, E., McCall, D. S., Ulanowski, Z., Stopford, C., and Mitchell, D. L., Mishra, S., and Lawson, R. P.: Cirrus clouds and cli- Kaye, P. H.: Application of RTDF to particles with curved sur- mate engineering: new findings on ice nucleation and theoretical faces, J. Quant. Spectrosc. Ra., 110, 1599–1603, 2009. basis, in: Planet Earth 2011, edited by: Carayannis, E., InTech, Hesse, E., Macke, A., Havemann, S., Baran, A. J., Ulanowski, Z., Rijeka, 257–288, 2011. and Kaye, P. H.: Modelling diffraction by facetted particles, J. Nelson, J.: Sublimation of ice crystals, J. Atmos. Sci., 55, 910–919, Quant. Spectrosc. Ra., 113, 342–347, 2012. 1998. Jolic, K. I., Nagarajah, C. R., and Thompson, W.: Non-contact, Neshyba, S. P., Lowen, B., Benning, M., Lawson, A., and optically based measurement of surface roughness of ceramics, Rowe, P. M.: Roughness metrics of prismatic facets of ice, J. Meas. Sci. Technol., 5, 671–684, 1994. Geophys. Res., 118, 3309–3318, 2013. Jourdan, O., Mioche, G., Garrett, T. J., Schwarzenböck, A., Vi- Nousiainen, T., Lindqvist, H., McFarquhar, G. M., and Um, J.: dot, J., Xie, Y., Shcherbakov, V., Yang, P., and Gayet, J. F.: Cou- Small irregular ice crystals in tropical cirrus, J. Atmos. Sci., 68, pling of the microphysical and optical properties of an Arctic 2614–2627, 2011. nimbostratus cloud during the ASTAR 2004 experiment: Im- Peterson, H., Bailey, M., and Hallett, J.: Ice particle growth under plications for light-scattering modeling, J. Geophys. Res., 115, conditions of the upper troposphere, Atmos. Res., 97, 446–449, D23206, doi:10.1029/2010JD014016, 2010. 2010. Kaye, P. H., Hirst, E., Greenaway, R. S., Ulanowski, Z., Hesse, E., Pfalzgraff, W. C., Hulscher, R. M., and Neshyba, S. P.: Scan- DeMott, P., Saunders, C., and Connolly, P.: Classifying atmo- ning electron microscopy and molecular dynamics of surfaces spheric ice crystals by spatial light scattering, Opt. Lett., 33, of growing and ablating hexagonal ice crystals, Atmos. Chem. 1545–1547, 2008. Phys., 10, 2927–2935, doi:10.5194/acp-10-2927-2010, 2010. Keramitsoglou, I., Harries, J. E., Colling, D. J., Barker, R. A., and Ruiz Gale, M. F., Hogert, E. N., and Gaggioli, N. G.: Apparent and Foot, J. S.: A study of the theory and operation of a resonance flu- real roughness, Opt. Laser. Eng., 45, 947–952, 2007. orescence water vapour sensor for upper tropospheric humidity Schön, R., Schnaiter, M., Ulanowski, Z., Benz, S., Möhler, O., measurements, Meteorol. Appl., 9, 443–453, 2002. Vogt, S., Wagner, R., and Schurath, U.: Particle habit imaging Kobayashi, T. and Ohtake, T.: Hexagonal twin prisms of ice, J. At- using incoherent light: a first step towards a novel instrument for mos. Sci., 31, 1377–1383, 1974. cloud microphysics, J. Atmos. Ocean. Tech., 28, 493–512, 2011. Korolev, A. V., Isaac, G. A., and Hallett, J.: Ice particle habits in Shcherbakov, V.: Why the 46◦ halo is seen far less often than the Arctic clouds, Geophys. Res. Lett., 26, 1299–1302, 1999. 22◦ halo?, J. Quant. Spectrosc. Ra., 124, 37–44, 2013. Korolev, A., Isaac, G. A., and Hallett, J.: Ice particle habits in strat- Shcherbakov, V., Gayet, J. F., Jourdan, O., Ström, J., and Minikin, iform clouds, Q. J. Roy. Meteor. Soc., 126, 2873–2902, 2000. A: Light scattering by single ice crystals of cirrus clouds, Geo- Krämer, M., Schiller, C., Afchine, A., Bauer, R., Gensch, I., Man- phys. Res. Lett., 33, L15809, doi:10.1029/2006GL026055, 2006. gold, A., Schlicht, S., Spelten, N., Sitnikov, N., Borrmann, S., Shilling, J. E., Tolbert, M. A., Toon, O. B., Jensen, E. J., de Reus, M., and Spichtinger, P.: Ice supersaturations and cir- Murray, B. J., and Bertram, A. K.: Measurements of the rus cloud crystal numbers, Atmos. Chem. Phys., 9, 3505–3522, vapor pressure of cubic ice and their implications for at- doi:10.5194/acp-9-3505-2009, 2009. mospheric ice clouds, Geophys. Res. Lett., 33, L17801, Kuhs, W. F., Sippel, C., Falenty, A., and Hansen, T. C.: Extent and doi:10.1029/2006GL026671, 2006. relevance of stacking disorder in “ice Ic”, P. Natl. Acad. Sci. Sölch, I. and Kärcher, B.: Process-oriented large-eddy simulations USA, 109, 21259–21264, 2012. of a midlatitude cirrus cloud system based on observations, Q. J. Lampert, A., Ehrlich, A., Dörnbrack, A., Jourdan, O., Gayet, J.- Roy. Meteor. Soc., 137, 374–393, 2011. F., Mioche, G., Shcherbakov, V., Ritter, C., and Wendisch, M.: Spichtinger, P. and Gierens, K. M.: Modelling of cirrus clouds Microphysical and radiative characterization of a subvisible mi- – Part 2: Competition of different nucleation mechanisms, At- dlevel Arctic ice cloud by airborne observations – a case study, mos. Chem. Phys., 9, 2319–2334, doi:10.5194/acp-9-2319-2009, Atmos. Chem. Phys., 9, 2647–2661, doi:10.5194/acp-9-2647- 2009. 2009, 2009. Takano, Y. and Liou, K. N.: Solar radiative transfer in cirrus Lu, R. S., Tian, G. Y., Gledhill, D., and Ward, S.: Grinding surface clouds. I. Single-scattering and optical properties of hexagonal roughness measurement based on the co-occurrence matrix of ice crystals, J. Atmos. Sci., 46, 3–19, 1989. speckle pattern texture, Appl. Optics, 45, 8839–8847, 2006. Ulanowski, Z.: Ice analog halos, Appl. Optics, 44, 5754–5758, Malkin, T. L., Murray, B. J., Brukhno, A. V., Anwar, J., and Salz- 2005. mann, C. G.: The structure of ice crystallised from supercooled Ulanowski, Z., Hesse, E., Kaye, P. H., Baran, A. J., and Chan- water, P. Natl. Acad. Sci. USA, 109, 1041–1045, 2012. drasekhar, R.: Scattering of light from atmospheric ice ana- McCall, D. S.: Measurement and modelling of light scattering by logues, J. Quant. Spectrosc. Ra., 79–80C, 1091–1102, 2003. small to medium size parameter airborne particles, Ph.D. thesis, Ulanowski, Z., Connolly, P., Flynn, M., Gallagher, M., University of Hertfordshire, Hatfield, 2011. Clarke, A. J. M., and Hesse, E.: Using ice crystal ana- Mishchenko, M. I. and Macke, A.: Incorporation of physical optics logues to validate cloud ice parameter retrievals from the effect and computation of the Legendre expansion for ray-tracing CPI ice spectrometer, in: Proc. 14 Int. Conf. Clouds Pre- phase functions involving d-function transmission, J. Geophys. cip., Bologna, 19–23 July 2004, 1175–1178, available at: Res., 103, 1799–1805, 1998.

www.atmos-chem-phys.net/14/1649/2014/ Atmos. Chem. Phys., 14, 1649–1662, 2014 1662 Z. Ulanowski et al.: Incidence of rough and irregular ice particles

uhra.herts.ac.uk/handle/2299/6753 (last access: 17 September Voigtländer, J., Bieligk, H., Niedermeier, D., Clauss, T., Chou, C., 2013), 2004. Ulanowski, Z., and Stratmann, F.: A new experimental setup to Ulanowski, Z., Hesse, E., Kaye, P. H., and Baran, A. J.: Light scat- investigate nucleation, dynamic growth and surface properties of tering by complex ice-analogue crystals, J. Quant. Spectrosc. single ice crystals, in: Geophys. Res. Abs., 15, EGU General As- Ra., 100, 382–392, 2006. sembly, Vienna, Austria, EGU2013-4206, 2013. Ulanowski, Z., Stopford, C., Hesse, E., Kaye, P. H., Hirst, E., and Yang, P. and Liou, K. N.: Single-scattering properties of complex Schnaiter, M.: Characterization of small ice crystals using fre- ice crystals in terrestrial atmosphere, Contr. Atmos. Phys., 71, quency analysis of azimuthal scattering patterns, in: Proc. 10 Int. 223–248, 1998. Conf. on Electromagnetic and Light Scattering, Bodrum, 17–23 Yang, P., Zhang, Z., Kattawar, G. W., Warren, S. G., Baum, B. A., June, 225–228, 2007. Huang, H. L., Hu, Y. X., Winker, D., and Iaquinta, J.: Effect of Ulanowski, Z., Hirst, E., Kaye, P. H., and Greenaway, R.: Retrieving cavities on the optical properties of bullet rosettes, Appl. Meteo- the size of particles with rough and complex surfaces from two- rol. Clim., 47, 2311–2330, 2008. dimensional scattering patterns, J. Quant. Spectrosc. Ra., 113, 2457–2464, 2012. Um, J. and McFarquhar, G. M.: Dependence of the single-scattering properties of small ice crystals on idealized shape models, At- mos. Chem. Phys., 11, 3159–3171, doi:10.5194/acp-11-3159- 2011, 2011.

Atmos. Chem. Phys., 14, 1649–1662, 2014 www.atmos-chem-phys.net/14/1649/2014/