The University of Manchester Research

Microphysical properties and radar polarimetric featureswithin a warm frontDOI:10.1175/MWR-D-18-0056.1

Document VersionAccepted author manuscript

Link to publication record in Manchester Research Explorer

Citation for published version (APA):Keppas, S., Crosier, J., Choularton, T., & Bower, K. (2018). Microphysical properties and radar polarimetricfeatures within a warm front. Monthly Weather Review, 146. https://doi.org/10.1175/MWR-D-18-0056.1

Published in:Monthly Weather Review

Citing this paperPlease note that where the full-text provided on Manchester Research Explorer is the Author Accepted Manuscriptor Proof version this may differ from the final Published version. If citing, it is advised that you check and use thepublisher's definitive version.

General rightsCopyright and moral rights for the publications made accessible in the Research Explorer are retained by theauthors and/or other copyright owners and it is a condition of accessing publications that users recognise andabide by the legal requirements associated with these rights.

Takedown policyIf you believe that this document breaches copyright please refer to the University of Manchester’s TakedownProcedures [http://man.ac.uk/04Y6Bo] or contact uml.scholarlycommunications@manchester.ac.uk providingrelevant details, so we can investigate your claim.

Download date:04. Aug. 2021

1

Microphysical properties and radar polarimetric features within a 1

warm front 2

S. CH. Keppas1, J. Crosier1,2, T.W. Choularton1, K.N. Bower1 3

1Centre for Atmospheric Sciences, SEES, University of Manchester, UK 4

2National Centre for Atmospheric Science, University of Manchester, Manchester, UK 5

6

Corresponding author e-mail: Jonathan Crosier, jonathan.crosier@manchester.ac.uk 7

8

Abstract 9

On 21 January 2009, the warm front of an extensive low-pressure system affected the UK weather. 10

In this work, macroscopic and microphysical characteristics of this warm front are investigated 11

using in-situ (optical array probes, temperatures sensors and radiosondes) and S-band polarimetric 12

radar data from the Aerosol Properties, PRocesses And InfluenceS on the Earth’s climate-Clouds 13

project. The warm front was associated with a warm conveyor belt, a zone of wind speeds of up 14

to 26m s-1, which played a key role in the formation of extensive mixed-phase cloud mass, 15

ascending significant liquid water (LWC ~0.22g m-3) at a level ~3km and creating an ideal 16

environment at temperatures ~ -5°C for ice multiplication. Then, “generating cells”, which formed 17

in the unstable and sheared layer above the warm conveyor belt, influenced the structure of the 18

stratiform cloud layer dividing it into two types of elongated and slanted ice fall-streaks, one being 19

depicted by large ZDR values and another by large ZH values. The different polarimetric 20

characteristics of these ice fall-streaks reveal their different microphysical properties, such as the 21

ice habit, concentration and size. We investigate their evolution, which was affected by the warm 22

conveyor belt, and their impact on the surface precipitation. 23

2

1 Introduction 24

Extratropical cyclones have been extensively studied, as such systems often demonstrate 25

complicated cloud structures providing an ideal environment for microphysical studies [e.g. 26

Matejka, 1980; Chen and Cotton, 1988; Forbes and Clark, 2003; Stark et al., 2013; Crosier et al., 27

2014; Lloyd et al., 2014; Dearden et al., 2016]. The typical structure of cyclones consists of acold, 28

a warm and an occluded front, which usually affect the UK with strong winds and heavy rainfall 29

[e.g. Browning, 2004; Lavers et al., 2011]. The fronts are associated with warm and cold conveyor 30

belts which circulate air masses with different traits. Warm conveyor belts (hereafter WCB), which 31

usually originate in surface maritime areas, convey humid air making a substantial contribution to 32

the cloud structure and surface precipitation [Harrold, 1973; Browning, 1986; Eckhardt et al., 33

2004; Pfahl et al., 2014]. Although all types of fronts have particularities and mechanisms that 34

need further investigation, studying warm fronts is important as they account for the majority of 35

the precipitation associated with extratropical cyclones [Keyser, 1986; Wakimoto and Bosart, 36

2001]. 37

Dual-polarization radars have been a valuable tool for revealing cloud microphysical 38

properties, providing details about ice crystal habits, dimension, shape, orientation and phase of 39

hydrometeors. Various studies have been conducted on linking polarimetric radar variable 40

thresholds with specific hydrometeor types [e.g. Straka, 2000; Hogan et al.,2002; Kennedy and 41

Rutledge, 2011; Schrom and Kumjian, 2016], and determining polarimetric signatures, which can 42

provide insight into cloud microphysical processes [e.g. Hall et al., 1984; Illingworth et al., 1987; 43

Hubbert et al., 1998; Smith et al., 1999; Ryzhkov et al., 2005; Kumjian and Ryzhkov, 2008; 44

Kumjian, 2013]. Comparing and linking in-situ cloud measurements with radar data can improve 45

and validate the interpretation of the remote sensing data, and therefore the understanding of cloud 46

microphysics, including the role of primary [Koop et al., 2000; Vali et al., 2015], secondary ice 47

production [Mossop and Hallett, 1974; Pruppacher and Schlamp, 1975; Vardiman, 1978; Knight, 48

3

1979 and Choularton et al., 1980] and embedded convection mechanisms. Embedded convective 49

turrets, which appear as cells with horizontal and vertical extent of up to 6km and 2km 50

correspondingly, are referred to as “Generating Cells” (GCs), and have been documented several 51

times in the literature [Wexler and Atlas, 1959; Houze et al., 1976; Hobbs and Locatelli, 1978; 52

Herzegh and Hobbs, 1981; Kumjian, 2014; Plummer et al., 2014; Rosenow et al., 2014; Plummer 53

et al., 2015; Oue et al., 2015; Rauber et al., 2015; Keeler et al., 2016a; Keeler et al., 2016b; Keeler 54

et al., 2017 and many others]. Discussing some of the latest studies, Plummer et al. [2014, 2015] 55

used in-situ (2D-C and 2D-P) together with a doppler radar (reflectivity factor and doppler 56

velocity) in order to investigate the microphysics of GCs and the fall-streaks originated in them. 57

In addition, Rosenow et al. [2014] used an airborne W-band radar to quantify the magnitude of the 58

vertical velocities in GCs, while Keeler et al. [2016a, b, 2017] used models to simulate GCs and 59

examine their origin, forcing and their relationship to vertical wind shear, ambient thermal 60

instability and cloud top radiative forcing. A mechanism that commonly triggers GCs is Kelvin-61

Helmholtz instability, which happens when velocity shear occurs in a single continuous fluid 62

[Browning et al., 1973; Hogan et al., 2002]. GCs usually present updrafts of <3m s-1, and they are 63

often associated with descending trails of ice crystals and snow, which are commonly referred to 64

as “ice fall-streaks”. Ice fall-streaks, the shape of which is affected by the vertical wind shear, 65

originate in and form from generating cells, play an important role in the production of 66

precipitation [e.g. Marshall 1953; Langleben 1956; Douglas et al., 1957; Wexler and Atlas, 1959; 67

Bader et al., 1987; Kumjian, 2014; Rosenow et al., 2014; Oue et al., 2015]. 68

In this study, we use high-resolution co-located in-situ and active remote sensing observations 69

obtained from dual-polarisation data, to provide a) detailed picture of the microphysical structure 70

of a warm front and b) analysis of the structure, the origin and the effects of the ice fall-streaks, 71

which are frequently embedded within warm fronts. 72

73

4

2 Methodology 74

The data used in this study were obtained during the NERC-funded APPRAISE-Clouds project 75

(part of the NERC Aerosol Properties, PRocesses And InfluenceS on the Earth’s climate 76

programme), which took place in the UK during 2007-2010. This cloud project aimed to provide 77

a comprehensive investigation of mixed-phase clouds influencing UK weather systems and the 78

impact of aerosols on the microphysical properties and development of such clouds [Crosier et al., 79

2011; Crawford et al., 2012]. 80

a. The in-situ datasets 81

The UK-FAAM (Facility for Airborne Atmospheric Measurements) Bae-146 aircraft was used 82

to collect in-situ measurements of cloud properties during APPRAISE-Clouds. In particular, on 83

21 January 2009 during flight B424, the aircraft extensively sampled clouds associated with a 84

warm front over southern England, performing a series of straight and level runs, profiles, and 85

saw-tooth profiles over the period 1500-2000UTC. The sampling occurred along a radial of 255o 86

within ~100km range from the Chilbolton Facility for Atmospheric and Radio Research (CFARR) 87

located at 51.15oN, 1.44oW, to allow direct comparison with radar scans along the same azimuth. 88

Hourly radiosondes were also launched from the Chilbolton ground site between 0700-1800UTC. 89

The FAAM Bae-146 aircraft was equipped with microphysical probes, which provide 90

measurements of a wide range of hydrometeors as described in Crosier et al., 2014. Key 91

instrumentation used in this work included: A mie scattering Cloud Droplet Probe (CDP) 92

measuring number and size (~3 and 50μm) of cloud droplets [Lance et al., 2013]; 2D-S (Two-93

Dimensional Stereo) and CIP-100 (Cloud Imaging Probe-100) Optical Array Probes (OAP) 94

providing shadow imagery of particles with size 10-1280μm and 100-6200μm [Knollenberg et al., 95

1970; Lawson et al., 2006] respectively; Rosemount probe measured the ambient de-iced 96

temperature. It should be highlighted that the 2D-S probe provides better resolution (10μm) and 97

5

wider size range (10-1280μm) comparing to its previous version 2D-C (25μm; 25-800μm). Particle 98

phase was estimated based on a shape analysis of images obtained with the 2D-S and CIP-100, 99

while ice number concentration and diameter (from the diagonal of the bounding box around the 100

recorded particle image) was estimated according to Crosier et al. [2011, 2014]. An inter-arrival 101

time threshold was used to filter OAP images to remove artefacts associated with shattering of 102

large particles on the probes according to Field et al. [2006]. 103

In order to estimate bulk ice-phase parameters, including mean diameter and concentration, 104

the 2D-S and CIP-100 data sets were merged, and measurements in overlapping sizes were 105

averaged. Ice Water Content (IWC) was calculated from merged 2D-S and CIP-100 particle size 106

distributions using estimates of particle mass from Heymsfield et al. [2007]. Significant 107

uncertainties (~50% for IWC) can arise because of the diversity in OAP image processing options. 108

Liquid Water Content (LWC) was estimated from the CDP particle size distribution. 109

b. The remote sensing datasets 110

During the flight, the Chilbolton Advanced Meteorological Radar (CAMRa) performed Range 111

Height Indicator (RHI) scans along the 255o (WSW) radial every ~90 seconds to obtain near co-112

located radar and in-situ measurements. The CAMRa is a dual-polarization Doppler radar that 113

operates at 3GHz with a steerable 25m dish resulting in a narrow 0.28o beam [Goddard et al., 114

1994]. Additionally, the zenith-pointing dual polarisation 35GHz Copernicus radar, which is 115

located within 50 meters of the CAMRa, was operating throughout the day. In contrast to low-116

frequency radars, which are applied for precipitation observations [Kollias et al., 2007; Fukao et 117

al., 2014], high-frequency radars are significantly affected by attenuation due to precipitation 118

[Godard, 1970]. The cross section ratio between large and small particles at such high frequencies 119

is smaller compared to longer wavelengths, so are an ideal tool for monitoring clouds and fog. All 120

references to data from the 35GHz vertical pointing radar will be indicated by the subscript “35”; 121

otherwise the data being discussed are from the 3GHz radar. 122

6

In this study, we use radar variables, such as reflectivity factor (ZH), Doppler velocity (VRAD) 123

and differential reflectivity (ZDR) to examine cloud dynamical and microphysical structures. ΖΗ 124

can be affected by the size and concentration of the hydrometeors [Doviak et al., 1979]. Doppler 125

velocity provides an estimate of the motion of hydrometeors, which can indirectly be used to infer, 126

depending on the mode of operation, information on wind fields or the fall-speeds of hydrometeors 127

[Browning and Wexler, 1968]. Finally, ZDR is the ratio between the radar reflectivities at horizontal 128

and vertical polarisations expressed in logarithmic scale induced by hydrometeors within the radar 129

beam [Seliga and Bringi, 1976], and can be used to provide details about their shape, orientation 130

and phase. We highlight that ZDR data coming from elevations <0.2o have been excluded due to 131

vertical beam blockage [Ryzhkov et al., 2002; Giangrande and Ryzhkov, 2005]. 132

133

3 The Synoptic Condition 134

On 21 January 2009, a deep low-pressure system in the northwest Atlantic Ocean moved 135

eastwards and developed (figure 1a). This system generated multiple frontal boundaries, which 136

approached the UK from the west. The Bae-146 research flight sampled the warm front as it passed 137

across the south of England. This warm front was observed to be associated with a WCB, which 138

originated in the warm sector of the system (figure 1). Figure 1a, b show that the WCB transported 139

humid and warm air over the colder air found over the UK mainland. In figure 1b, the WCB 140

becomes perceptible through the zone of enhanced relative humidity and strong SSW-W winds 141

(up to 23m s-1), which reached a level of ~450mb (~6km height). 142

143

144

4. A Macroscopic Description of the cloud features 145

7

A brief description of the entire in-situ and radar dataset is provided in order to understand the 146

general cloud structures in the lower/middle troposphere during the passage of the warm front 147

shown in figure 1a. 148

The spatial structure of the frontal cloud obtained from CAMRa at 1712UTC is depicted in 149

figure 2. Several features, which can be seen in figure 2, were consistently observed throughout 150

the frontal passage and will now be discussed in turn. Firstly, we observed GCs presented as 151

protruded cloud tops, which demonstrated high-ZH/low-ZDR cores (>10dBZ / ~0dB) and low-152

ZH/high-ZDR (<10dBZ / >1dB) boundaries (figure 2a, d) [e.g. Bader et al., 1987; Kumjian, 2014; 153

Oue et al., 2015]. A ZH ice fall-streak [e.g. Bader et al., 1987; Kumjian, 2014; Oue et al., 2015] is 154

represented as a long (tens of km) and narrow (~1km) slant (angle of 3-7o) zone of high ZH 155

(enhanced by 10-15dBZ relative to the surroundings) (figure 2a) and ~0dB ZDR (figure 2d) 156

containing mostly rimed/aggregated crystals (figure 5, 2km<altitude<3.5km). Then, ZDR fall-157

streaks [e.g. Bader et al., 1987; Kumjian, 2014; Oue et al., 2015] typically located above a ZH fall-158

streak, exhibited a zone of high-ZDR (>1dB) (figure 2d) and low-ZH (<10dBZ) (figure 2a) values 159

consistent with the presence of pristine ice crystals (figure 5, 3km<altitude<5km). Although 160

figure 2 does not show an ideal example of GC, it highlights the connection between a GC and a 161

ZH/ZDR fall-streak (see details in section 5b). Finally, figure 2f presents two banded layers of 162

enhanced Doppler velocity associated with the so called Warm Conveyor Belt (hereafter WCB), 163

an air stream which appears in RHIs between 2-4km, which transports humid and warm air aloft. 164

An overview of the major temporal changes in the cloud structure can be seen in the vertical-165

pointing 35 GHz radar data (figure 3). It should be noted that clouds located at horizontal distances 166

between 30-90km range in the 3GHz radar scans are estimated to overpass the 35GHz radar 25-167

75 minutes later (based on a typical wind speed of ~20m s-1 along the radar profile). Although 168

comparisons between these two radar datasets will not be fully correlated as the system may 169

change during transit, it reveals useful general trends of the system over time. 170

8

In figure 3a, we can see a change in the height of the freezing level or “melting layer” (referred 171

to as ML). The ML appears as a bright band in the reflectivity field, the ZH signal of which 172

increases depending on the type and concentration of the hydrometeors that enter this region 173

[Fabry and Zawadski, 1995]. ZH increases within the ML due to the change of the phase and shape 174

of melting hydrometeors, which also affects ZDR [e.g. Stewart et al., 1984; Zrnic et al., 1993; 175

Szyrmer and Zawadski, 1999; Giangrande et al., 2008]. Before the passage of the warm front 176

(<2215UTC) the ML was observed at ~0.7-1km (ZH_35 reached 30dBZ at 0.75km; figure 3a). After 177

the passage of the warm front, the ML height raised to ~1.6km (2215 UTC) (figures 1a, 3a). The 178

ML is also depicted as a zone of dramatic increase of Doppler velocity (from ~1 to >4m s-1) 179

retrieved by the 35 GHz vertical-pointing radar (figure 3b), because hydrometeors may become 180

larger in size due to coalescence and denser due to melting [Mitchell, 1996; Heymsfield and 181

Iaquinta, 2000; Protat and Williams, 2011]. 182

Data from the 35 GHz vertical-pointing radar (figure 3a) show that the cloud top height 183

decreased (from 6.5km to 5km) at a mean rate of 0.24km/hour during the frontal passage, while 184

the cloud top temperature increased from ~-35°C to ~-25°C according to radiosondes (figure 4). 185

The lowering of cloud top height with time, which can be seen in figure 4c as a decrease of RH 186

with time by 50% at altitudes 6-7km, is also perceptible from figures 2b (3GHz radar) and 3b 187

(35GHz radar). The 35GHz radar data (figure 3c) show that >50% of the pixels were considered 188

as cloudy (where ZH>-40dBZ) at altitudes ≤5.5km for 1800-1900UTC. However, this height 189

gradually decreased to 4.5km (1900-2000UTC) and 4km (2100-2200UTC). The corresponding 190

heights for the 3GHz radar were always observed ~0.5km lower than the 35GHz radar (figure 2b), 191

as the 3GHz frequency is mainly used for the detection of precipitation and not of smaller cloud 192

particles. According to figure 3b, Doppler velocities between 0 and -1m s-1 were observed at cloud 193

tops (height >4km), which is consistent with weak vertical motions and slowly precipitating low 194

density ice crystals (like the small ice-plates and columns at altitudes >4km and temperatures <-195

15°C in figure 5) [Jayaweera and Cottis, 1969; Heymsfield, 1972; Kajikawa, 1972]. 196

9

In order to gain insight into the presence of the aforementioned features, a statistical analysis 197

of the ZH and ZDR datasets is shown in figures 2c, e and 3c. In relation to ZH fall-streaks, in figure 198

2c, ZH seems to increase through the time between 1646-1932UTC (95th percentiles of ZH ~20dBZ 199

at 2-4km height). At 1912-1932UTC, although average ZH decreased significantly to ~10dBZ, 200

high ZH 95th percentile values (>20dBZ) were observed at 2-4km, highlighting the presence of 201

discrete GCs. This is also demonstrated by the 35GHz radar dataset, as 95th percentiles of ZH-35 202

around the ML were enhanced at 2000UTC (from 15dBZ to ~30dBZ). This ZH-35 increase, but also 203

the enhanced 95th percentiles at 2-4km, implies the occurrence of more frequent and more intense, 204

but spatially restricted GCs, which occasionally affected the ML echo (figure 3d). This is also 205

demonstrated by the fact that ZH-35 increased in the ML and decreased at 2-4km with time, while 206

95th percentiles signals at 2-4km remained significantly high (~15dBZ). In addition, the vertical 207

pointing-radar at 1900-2030UTC observed larger Doppler velocities (-2 to -3m s-1) comparing to 208

the previous time period at 0.7-3km height, which indicate the existence of denser hydrometeors 209

(like the large ice particles at altitudes <3.5km in figure 5). 210

Switching focus to ZDR fall-streaks, these features can be observed in figure 2e as local ZDR 211

maxima at altitudes of ~4km (slightly higher than ZH fall-streaks). The ZDR fall-streaks manifest 212

mainly as perturbations to the 95th percentile (0.5-2dB) relative to the rest of the profile affecting 213

also the mean value (0.5-1dB). The ZDR fall-streaks are mainly observed during the second half of 214

the observing period, 1800-1932UTC. 215

The WCB was a clearly identifiable dynamical feature which likely “fed” the cloud system 216

with water vapor. It is presented, in the vertical-pointing radar dataset, as a zone of Doppler 217

velocity ~0m s-1, where strong westerlies (typically >20m s-1) of the WCB advected the 218

hydrometeors almost horizontally (figure 3b). Whilst the wind speed profiles suggest significant 219

stratification at all times (figure 4d), it seems that the highest wind speeds were observed near 220

cloud top, where relative humidity (figure 4c) and cloudiness (figure 3c) dramatically decrease. 221

10

The main region of high wind speeds (20-30m s-1) (figure 4d) and positive wind shear (figure 2h) 222

was firstly located at 5-6km (1500-1600) but gradually lowered to 3.5-4.5km (after 18:00UTC). 223

However, before 1800UTC, there were also other enhanced wind speed layers along the profile, 224

which likely represent slight differences in air mass characteristics/origin. As an illustration, two 225

well-distinguished layers centered at altitudes of ~2.5km and ~4km at ~1700UTC (figure 2f), 226

coincided with wind speed (17-20m s-1), relative humidity (>90%), clock-wise wind veering (~60o 227

km-1 comparing to the average of ~20o km-1 for altitudes 0-6km) and potential temperature (~ +9K 228

km-1 comparing to average rate of +4K km-1 for altitudes 0-6km) maxima (figures 4c, d, e, f). At 229

the same altitudes, temperature presented the highest increase (up to 1.3oC h-1) between 1500-230

1800UTC. All the above observations reveal that there was a main warm front at ~4km and a 231

possible sub-frontal zone at ~2.5km, the locations of which were associated with defined structures 232

in the wind speed and temperature data. As a general trend, it seems that as the warm front was 233

approaching, the various wind speed layers tended to merge (only one significant wind speed peak 234

can be observed at ~4km at 1800UTC – figure 4d) and were located at lower altitudes (figures 235

2g, h). 236

237

5 Microphysical properties of the warm front 238

Using in-situ and remote sensing data, the microphysical properties in specific features of the 239

warm frontal cloud, such as the WCB, ice fall-streaks, and regions of secondary ice production are 240

investigated. 241

a. The Warm Conveyor Belt (WCB) 242

The WCB originated in the warm sector of the deep low-pressure system shown in figure 1a 243

and, during the observing period, it intersected the surface in the Celtic sea. It was responsible for 244

11

the large-scale slantwise ascent of humid air, which led to widespread formation of mixed-phase 245

clouds. 246

At 1600UTC, CAMRa scans to the WSW identified three zones of enhanced Doppler Velocity 247

(up to ~20m s-1) at altitudes of 2.5km, 4km and 6km respectively (figure 6a). These three zones 248

caused significant vertical wind shear (up to -12m s-1 km-1), which could potentially release 249

instability and form convective GCs. Similar regions of wind shear were also identified in the wind 250

speed and direction data from radiosonde profiles (figure 6b). In regions where the radar scan 251

azimuth (255o) is closely aligned with the wind speed, (at altitudes 2.5-3km), the radiosonde wind 252

speed is in good agreement with Doppler velocity. 253

At 1800UTC, as the warm front approached the radar site, a broad slanted zone of high Doppler 254

velocities (up to 23m s-1) was observed spanning altitudes of 2-5km (figure 6c). The high Doppler 255

velocity zone above 3-5km highlighted the WCB location and presented enhanced wind speed 256

(15.6-26.0 m s-1) and wind direction shear (21o km-1). The peak wind speed was located near the 257

middle of the WCB, with two zones of wind speed and direction shear located below (+8.7m s-1 258

km-1, ~47o km-1) and above it (-3.8m s-1 km-1, ~+10o km-1). Again, there was a good level of 259

agreement between Doppler velocity and measured wind speed for wind directions close to the 260

radar beam azimuth (255o) (figure 6a-d). Comparing and linking the Doppler velocity with 261

radiosonde data, we estimate the main warm front location is as shown in figure 6c by the solid 262

red line. Smaller fluctuations in Doppler velocity fields and radiosonde wind profiles indicate the 263

presence of a small “sub-frontal” zone located at ~2km (dashed red line). 264

At 1935UTC, the frontal system had transited further to the east over the ground site. At this 265

time the WCB was located ~2km closer to the surface (between 1-3km), being represented by 266

distinct zone of Doppler velocities between 17-25m s-1 (figure 6e). Aircraft observations obtained 267

through the WCB at this time indicate the presence of significant liquid water (figure 6f) (up to 268

0.22g m-3) and cloud droplet concentrations (10-58cm-3), while IWC dramatically increases 269

12

outside the WCB (up to 0.044g m-3). These observations support the idea that slantwise ascent 270

from the WCB generates a large expanse of cloud containing super-cooled liquid water. The 271

lifetime and radiative properties of similar super-cooled layer clouds is sensitive to the presence 272

of Ice Nuclei [e.g. Pinto, 1998; Jiang et al., 2000; Morrison et al., 2005; Murray et al., 2012], 273

which can activate to form ice in the cloud leading to poor representation in weather models. 274

275

b. The Generating Cells (GCs) 276

As the warm front approached the UK, some GCs appeared in RHIs, especially after 1800UTC. 277

Despite the lack of in-situ measurements within these features, we try to investigate their 278

microphysics using the available radar data. In general, GCs were observed above the WCB and 279

above a braided structure [Chapman and Browning, 1998] associated with sheared wind (figure 280

7a). Sheared wind is associated with Kelvin-Helmholtz instability [Browning, 1971; Browning et 281

al., 1973], which is a mechanism that can trigger convection [Hogan et al., 2002]. 282

At altitudes where GCs appeared (4.5-6.5km), the rate of change in potential temperature with 283

altitude recorded by the radiosondes at 1700 and 1800UTC was 0-5K km-1 (figure 4f). It is 284

important that a small region of negative rate is observed at ~5km, which coincides with the typical 285

location of GCs, implying that GCs formed due to the release of instability through the Kelvin-286

Helmholtz mechanism. In the meantime, the wind speed exhibited an increasing trend with time 287

within the WCB (4.2km) (figure 4d). This increased the wind shear from ~4 to ~8m s-1 km-1 at 288

1700 and 1900UTC respectively at the top layer of the WCB, which also lowered from 4.5 to 289

3.5km (figure 2h). Although a radiosonde cannot provide representative data over a wide area, the 290

above measurements indicate the occurrence of instability and a triggering mechanism which can 291

explain the presence of GCs. As stated earlier (beginning of section 4), GCs demonstrated a core 292

of ZH >10dBZ, which increased to 20-33dBZ as the warm front approached the Chilbolton site. 293

13

GCs also demonstrated high ZDR at cloud tops (>2dB), which indicate regions of newly formed 294

ice, in which ZDR fall-streaks originated (e.g. figure 7e, range=85-95km). 295

In figures 7a-h, we present the evolution of a GC. At ~1928UTC (figures 7a, e), a GC was 296

triggered at approximately 90km distance from CFARR, above the WCB. The Doppler velocity 297

below the GC was ~ -23m s-1, but only ~ -14m s-1 in the newly formed GC, indicating wind shear 298

at the top of the WCB (figure 7a). The above observations suggest that the trigger which caused 299

the formation of GCs was the wind shear at the WCB top layer (Kelvin-Helmholtz instability). 300

Through this mechanism, significant amounts of liquid water from the WCB are lofted via 301

turbulence and weak updrafts. In the example presented in figures 7i, j, k, the aircraft measured 302

high cloud droplet concentration and LWC (up to 10cm-3 and 0.3g m-3 correspondingly) at the rear 303

region of a newly formed GC (located within the WCB at 55km<range<65km, 304

2km<altitude<3km). It is important to highlight that no positive values of Doppler velocity were 305

recorded by the vertical-pointing radar (figure 4d). According to the CAMRa, GCs were generally 306

forming at ranges >30km away, passing overhead the vertical-pointing radar as dissipated cells or 307

ZDR/ZH fall-streaks. In addition, due to short time of GC genesis (5-15minutes) and their narrow 308

updraft region (typically <5km), it is not highly likely that a GC was captured overhead the radar 309

during its genesis time. 310

The GC was further developed in height 10-20 minutes later (figures 7b, f and c, g), which 311

caused the intensification of the aggregation and riming processes and affected the ZH parameter 312

(ZH increased to ~32dBZ). At the later stage (figure 7g), a distinct region of very high ZDR and 313

low-medium ZH was observed (up to 4dB, <17dBZ) at the GC top. This suggests regions of newly 314

formed (as there was not such a region earlier) pristine ice-plates/dendrites at -15°C (figure 7f) 315

[e.g. Magono and Lee, 1966; Schrom et al., 2015; Bailey and Hallett, 2009; Schrom and Kumjian, 316

2016]. The fact that these ZDR structures, which indicate the presence of un-rimed ice particles, 317

typically form at the latter stages of GC formation, suggests that they result from ice nucleation in 318

regions where the initial convective feature has mostly decayed, but that significant regions of 319

14

supersaturation still remain. Finally, as the GC core sediments/precipitates into the WCB (figure 320

7d, h), the associated ZH structure changes from a largely vertical orientation, to being more 321

horizontally elongated due to strong westerlies within the WCB (figures 14d, i). The evolution of 322

this GC shows the connection between fall-streaks and GCs. Another example is also shown in 323

figures 12c, f (40km<range<60km, 2km<altitude<4km) 324

325

c. The ZDR and ZH ice fall-streaks 326

(i) General Characteristics 327

At 1800UTC, two zones with different radar polarimetric traits were detected within sheared 328

ice fall-streaks. ZDR fall-streaks appeared as slanted zones of significantly high ZDR (>1.5dB) and 329

low ZH (<15dBZ) values, containing pristine ice crystals (hexagonal plates and dendrites) [e.g. 330

Andric et al., 2013; Schrom et al., 2015; Schrom and Kumjian, 2016]. ZH fall-streaks, which 331

appeared as a zero ZDR (-0.5 to 1dB) and high ZH (>15dBZ) zone, were typically observed beneath 332

a related ZDR fall-streak containing mostly aggregated and rimed ice crystals. Essentially, these 333

fall-streaks were represented by a bipolar zone of high-low ZDR (or low-high ZH) zones (similar 334

fall-streaks have been investigated by Bader et al. [1987], Kumjian [2014], Oue et al. [2015]). 335

These zones originate in and descend from GCs (section 5b), as ice crystals fall into sheared flow 336

[Kumjian, 2014]. ZH fall-streaks can be also enhanced by pristine ice crystals precipitating into it 337

from the overlying ZDR fall-streak. 338

Examining the entire in-situ and polarimetric radar dataset, observations of ZDR >1.5dB were 339

mostly (>60% of the observations) linked with low ice concentrations (<2L-1) and small-sized ice 340

particles (<800μm) (figure 8a, b). Taking into account previous literature [e.g. Schrom et al., 341

2015; Schrom and Kumjian, 2016] and the fact that the data around the ML (<1km) were excluded, 342

it seems that large ZDR values are mainly linked with small pristine ice crystals in small 343

concentrations within the ZDR fall-streaks. Figure 8c also shows that regions of ZDR<1dB are 344

15

partially associated (by >45%) with observations of ZH >15dBz due to ZDR fall-streaks, while 345

regions of ZDR >1dB are associated (by 60-70%) with smaller ZH (ZH <15dBz) due to ZH fall-346

streaks. It should be highlighted that regions of lower ZDR values (in general <1.5dB such as ZH 347

fall-streaks) were observed to have a larger variety of ice particle number concentrations (from 348

<4L-1 to 10L-1) and mean size (from <500 to 3600μm). Smaller crystals in larger concentrations 349

can be explained by the effects of secondary ice production (see section 5d). To conclude, although 350

the ZH/ZDR fall-streak radar polarimetric boundary is not particularly clear, a rough threshold for 351

distinguishing them could be: ZH ~15dBZ, ZDR~1dB. 352

An example of the evolution of large scale, moderately intense ZDR/ΖΗ fall-streaks is illustrated 353

in figure 9. The first fall-streak started to form at ~1549UTC (figures 9a, e), being located at 354

range=100-110km and height=4-6km. In this region, a spatially more extensive but less intense, 355

in terms of ZH (ZH <17dBZ), GC occurred (compared to the GC investigated in section 5b). At 356

this stage, the ZDR fall-streak was relatively modest, occasionally approaching 1-2.5dB. The 357

corresponding ZH structure was also relatively modest, exhibiting values of ~15dBZ, which was 358

an enhancement of ~10dBZ over the ZDR fall-streak. As the wind speed in the WCB appears to 359

increase ~25 minutes later (16:13 UTC, figures 9b, f), both ZDR and ZH fall-streak signals become 360

more obviously enhanced (up to 2.7dB and 24.5dBZ respectively). The ZH fall-streak coincided 361

with a region of ZDR~0dB, and is first observed in a region of enhanced Doppler velocity (>20m 362

s-1). As this region was part of the WCB (which transported large amounts of liquid water), it is 363

highly likely that ice crystals became intensely rimed and less oblate within it. Ten minutes later 364

(~16:23, figures 9c, g), the ZH fall-streak was represented by -0.6dB<ZDR<1.5dB and ZH >15dBZ, 365

containing large (~984μm) aggregated and heavily rimed dendrites (figure 9i) in concentrations 366

of ~1.4L-1 (figure 9m). The ZH fall-streak was an almost glaciated zone typically exhibiting LWC 367

~0g m-3 and IWC >0.01g m-3. 368

16

After a further 30 minutes (~16:55), the fall-streak was fully developed demonstrating the 369

largest length of all the observed fall-streaks (~60km) during the day, and an expansion rate of 370

~17m s-1. Although the ZDR fall-streak was characterised by ZH<15dBZ and ZDR>1dB due to small 371

concentrations of ice-dendrites and plates (figures 9d, h), it was actually divided into two sub-372

fall-streaks: one (figure 9d, h; range=65-85km, ZDR >1.5dB, T~ -13°C) with small (~580μm) ice-373

plates (fig. 9j – frames 3-4) and another (figure 9d, h; range=55-65km, ZDR ~1.5dB, T~ -13.5°C) 374

with large (400-1400μm) ice-dendrites/dendrite aggregates (fig. 9j – first two frames; 9n), both 375

in small concentrations (<1 L-1 and <5L-1 respectively). We speculate that the sub-fall-streak which 376

contained ice-dendrites, formed when small and light ice-plates drifted away from the initial ZDR 377

fall-streak due to strong westerlies and remained within the WCB for a longer time. As a result, 378

these plates grew into ice-dendrites due to the high supersaturation regime at ~ -13oC (figure 9d) 379

[Bailey and Hallet, 2009]. In the transitional zone between the ZDR and ZH fall-streak; (figure 9d, 380

h; range=90km, height=3.5km), small graupel particles mixed with slightly larger pristine plates 381

were observed in small sizes (~460μm), but in larger concentrations (up to 8L-1) (fig. 9j – last two 382

frames). In total, although this enhanced ZDR fall-streak reached the ML retaining its 383

characteristics, there was a decrease of ZDR slightly above the ML, presumably due to aggregation. 384

An important conclusion that arises from both sections 5b and 5c(i) is that the slope of the fall-385

streaks increased as the warm front was approaching. In particular, the slope of the ZDR fall-streak 386

at 1927UTC (figures 7a, e) was ~5o comparing to ~3o and ~7o at earlier (16:56UTC) and later 387

(20:07, figure 7d) times. It seems that the wind speed intensity of the WCB (discussed in sections 388

4 and 5a) and, thus, the wind shear intensity at its upper boundary, played an important role in the 389

formation and shape of the fall-streaks. Thus, stronger wind shear might cause stronger updrafts 390

forming larger and heavier hydrometeors in a shorter period of time, which fell faster towards the 391

surface. 392

393

17

(ii) The impact of the ZH fall-streak on the surface precipitation 394

The identification of the ZH fall-streak is very important as it is directly related to precipitation 395

enhancement at the surface. In figure 10, we present the evolution of the ZH/ZDR fall-streak dipole, 396

which was described in section 5c(i), with the corresponding rain rate product from the UK Met 397

Office operational radar network (NIMROD). 398

The ZH fall-streak remained elongated (length ~60km) for ~30mins (1656-1717 UTC) before 399

dissipating, exhibiting up to ZH ~28dBZ. In the initial stages (1705UTC; (figures 10a, d) it 400

appeared to be moving slantwise downwards, being represented by ZH_3GHz >20dBZ and ZDR~0dB. 401

As the ZH fall-streak evolved over the next hour, ice crystals moved slantwise downwards, with 402

aggregation and riming leading to greater and greater enhancements of the ZH signal. At 1804UTC, 403

a region of significantly high ZH (30-47dBZ) was observed at the ML (~0.7km). Such enhanced 404

ZH values imply heavy aggregation due to higher temperatures (>-3.5oC; figure 2d) and, thus, 405

“stickier” ice crystals. Comparing the surface rain rate graphs (figures 10i-l) with the RHIs, it 406

seems that the surface precipitation is strongly affected by the ZH fall-streak. In particular, the rain 407

rate increased (at range=30km; fig. 10i-l) from <1mm h-1 to 5mm h-1. This suggests that fall-streaks 408

which are initiated near cloud top, and develop over time scales of approximately an hour, have a 409

significant impact on surface precipitation rates downstream. 410

Finally, although a co-polar correlation coefficient (ρHV) was not available in order to estimate 411

the ML height [e.g. Giangrande 2008; Boodoo et al., 2010], ZDR (2-4dB) peaked around the ML 412

at different heights in (range=32km) and out (range=25km) of the ZH fall-streak (figure 11). In 413

particular, at 1800UTC, the ML was located at 0.8km (figure 4a), which broadly agrees with the 414

ZDR/ZH peak [Brandes and Kyoko, 2004; Houze, 2014, pg.144] of the 25km vertical profiles 415

(figure 11, green lines). However, a slight lowering of the ZDR peak height was observed within 416

the enhanced ZH region (by 0.2-0.4km). The local depression of the melting layer was possibly 417

caused by the melting of large aggregates [Stewart,1984; Stewart et al., 1984; Oraltay and Hallett 418

18

2005; Griffin et al., 2013]. This is important as ~0oC isothermal layers can be produced close to 419

the surface allowing (partially melted) ice to precipitate [Findeisen, 1940; Szeto et al., 1988; 420

Ryzhkov et al., 2011; Griffin et al., 2013]. 421

d. Secondary ice in the ZH fall-streak 422

In this section, we try to analyse concurrent in-situ and radar data in order to detect, determine 423

and investigate the characteristics of regions, which appear to be greatly influenced by secondary 424

ice processes. 425

Between 19:12:54-19:17:42, the aircraft passed through a region (range=50-70km) of low-426

medium ZH (8-12dB) and low ZDR (0.5-1dB), which was located at the entrance of the WCB. 427

Although this region was located within a weak signal cloud top, we speculate that this region is a 428

mature/dissipated ZH fall-streak. At the rear of this mature ZH fall-streak (range=65-80km), which 429

was located within the WCB entrance (Doppler velocity >20m s-1), the aircraft measured enhanced 430

LWC and cloud droplet concentration (up to 0.17g m-3 and 16.7cm-3 respectively). In contrast, at 431

the bottom of the WCB (range=50-65km), both cloud droplet concentration and LWC decreased 432

by 10cm-3 and 0.10g m-3 respectively. This liquid water removal may imply intense riming, which 433

is a mechanism assisting in ice multiplication [Mossop and Hallett, 1974; Choularton et al., 1978; 434

Cloularton et al., 1980]. The occurrence of small (~300μm in average) ice-columns (figure 12g; 435

2D-S frames 1-4) in high ice number concentrations (up to 37.8L-1) and the enhanced IWC (0.11g 436

m-3) observed at temperatures ~ -4.8oC imply that Hallett-Mossop is the possible ice multiplication 437

mechanism here. Some aggregated ice-columns were observed at range=65-80km (figure 12g, 438

CIP-100 frames 1-2) formed due to the occurrence of ice-columns within a region of high LWC. 439

Further to the east (figure 12c, f, h, i), a region of quasi-spherical rimed ice particles (figure 440

12g; 2DS frames 4-5, CIP-100 frame 3) and supercooled water drops mixed with some ice-441

columns (range=43-55km; T ~ -4°C; ZH up to 28.6dB and ΖDR~ 0.5-1dB) was observed, followed 442

by another region of mostly ice-columns (range=31-43km; T ~ -5°C). As this region belonged to 443

19

a short and newly formed ZH fall-streak originating in the core of a GC (range ~53km, altitude 444

~3km), processes, such as riming and ice-multiplication could be at early stages. This may explain 445

the increased, but lower than the previous secondary ice region, ice concentrations (up to 14.1L-446

1). Decreased LWC (<0.005g m-3) and cloud droplet concentration (<0.4cm-3) could be measured 447

due to riming process and the aircraft position (outside the WCB). It should be noted that some 448

pristine ice crystals from the ZDR fall-streak above (range=40-50km) rimed (an example in figure 449

12; CIP-100 frame 4) and became heavier, when they moved into the WCB. As a result, they 450

might fall into the ZH fall-streak, which was finally characterized by small quasi-spherical (and 451

possibly mixed-phase) ice particles/supercooled water drops, but also of ice-columns and ice-452

lollies [Keppas et al., 2017]. 453

454

6 Conclusions 455

On 21 January 2009, multiple frontal zones affected the UK weather due to a maturing 456

depression, which originally formed over the North Atlantic Ocean. In the work presented, the 457

dynamics and microphysics of mixed-phase clouds, with embedded convective elements, 458

associated with the warm front of this system are investigated comparing high resolution (0.28° 459

beam) dual-polarisation S-band radar with in-situ aircraft data. The latter data collected by 2D-S 460

probe (providing better resolution than 2D-C), which were used together with CIP-100 probe data 461

in order to provide a wider size range of measured particles. It should be noted that this is the first 462

time that a warm front is comprehensively investigated with high resolution data. 463

A few studies previously investigated warm fronts and GCs comparing both in-situ and radar 464

datasets. Here we discuss some of their results, which generally come to an agreement with the 465

results of the present study. Herzegh and Peter [1980] found that large ice particles within GCs 466

grew by riming, deposition and aggregation as occurred in ZH fall streaks. They also suggested 467

that GCs work as feeders of ice to stratiform clouds below. According to Matejka et al. [1980], as 468

20

GCs mature, they tend to be glaciated. They supported that surface precipitation is mainly 469

associated with embedded convection in warm fronts. Hogan et al. [2002] demonstrated some 470

evidence that embedded convection within a warm front is triggered by Kelvin-Helmholtz 471

instability. However, they found that embedded convection was linked with narrow vertical high 472

ZDR zones of ice columns, which coincided with updrafts and was a feature that was not observed 473

in the present study. Murakami et al. [1992] observed only a shallow layer of supercooled droplets 474

above the warm frontal zone. Additionally, they found that GCs provided a favorable environment 475

for ice crystals to grow rapidly, which agrees with Plummer et al. [2014]. Finally, Plummer et al. 476

[2015] noticed that ice mainly grew in regions below GCs, where enhanced moisture was 477

observed. 478

As a conclusion, the general structure of the warm front is schematically summarised in figure 479

13, according to the airborne in-situ and ground-based polarimetric radar measurements. The main 480

conclusions of the study are outlined as follows: 481

Regarding the macroscopic characteristics of the investigated warm front: 482

The height of cloud tops gradually decreased from 6km to 4.5km as the warm front was 483

approaching. Cloud top temperatures were, generally, >-25oC, which implies the ice in 484

the cloud system was formed via primary heterogeneous processes, and potentially 485

enhanced by secondary multiplication processes. 486

The height of the 0oC level, or the melting layer (ML), was firstly located at 0.7km 487

rising by almost 1km after the passage of the warm front. 488

The horizontal dimension of the cloud mass was larger than 120km, the maximum 489

range of the RHIs presented in this work. 490

The cloud mass consisted of distinctive features which demonstrated definite radar 491

polarimetric and/or microphysical characteristics. These features are: 492

21

the Warm Conveyor Belt (WCB), was initially depicted by a group of multiple zones 493

of enhanced Doppler velocity. These multiple zones merged into a single one, which 494

decreased in altitude, as the warm front moved further inland away from the coast. High 495

radar Doppler velocities (20-30m s-1) and radiosonde wind speeds (up to 26m s-1) were 496

recorded within the WCB presenting, essentially, a fair agreement for wind directions 497

along the radar azimuth. The WCB was vitally important for transporting large amounts 498

of liquid water (up to 0.37g m-3) into the system, offering at the same time an ideal 499

regime for secondary ice production at temperatures ~ -6oC. 500

the Generating Cells (GCs), formed at the unstable, based on potential temperature 501

profiles, layer above the WCB, where vertical wind shear triggered Kelvin-Helmholtz 502

instability. The GCs were represented by a core of high-ZH (10-33dBZ)/~0dB ZDR and 503

a shell of low-ZH (<10dBZ)/high-ZDR (up to 4dB). 504

the ice fall-streaks, formed as GCs were moving into a sheared flow caused by the 505

WCB. The strong westerlies “converted” the GCs’ high-ZDR shells and high-ZH cores 506

into an ice fall-streak, which consisted of two individual fall-streaks with different 507

polarimetric characteristics: a ZDR and a ZH fall-streak respectively. Ice fall-streaks 508

exhibited lengths of up to 60km and slope of 3o, which increased with the time (up to 509

7o) due to the intensification of the GCs. The ZDR fall-streaks were divided into two 510

low ZH (<15dBZ) subzones: a.) a zone of ice-plates with ZDR>1.5dB and b.) a zone of 511

dendrites with ZDR~1.5dB, which might form from small ice-plates that were moving 512

from GC shells to the high supersaturated regions of the WCB. As icing can be harmful 513

for aviation in dendritic ice regions, this can help to improve existing icing detection 514

algorithms [Serke et al., 2008; Pitersev and Yanovsky, 2011; Ellis et al., 2012]. Beneath 515

the ZDR fall-streaks, ZH fall-streaks were represented by high-ZH (>15dBZ) and 516

ZDR~0dB being consisted of mostly rimed and/or aggregated ice crystals in variable 517

sizes (from <500μm to 3600μm) and concentrations (from <4L-1 to >10L-1). ZH fall-518

22

streaks that form higher up in the clouds can affect the surface precipitation. The time 519

and spatial evolution depends on the profile of wind direction and speed in the clouds. 520

As an illustration, a case where a ZH fall-streak enhanced the surface precipitation by 521

4mm h-1 an hour after its genesis was presented. 522

the secondary ice regions, which presented large ice concentrations (up to 37.8L-1) 523

and mostly consisted of ice-columns. Such regions were located within the intersection 524

region of the ZH fall-streaks with the WCB, where riming was intense at temperatures 525

~ -5oC and liquid water was dramatically depleted potentially affecting the cloud 526

lifetime. 527

In the present paper, both heterogeneous ice nucleation and secondary ice formation were 528

observed in mixed-phase clouds, which are highly uncertain processes. Heterogeneous ice 529

nucleation plays an important role in the formation of GCs, as new ice forms from freezing liquid 530

water transported aloft along the rear updraft region of GCs. As in-situ measurements within GCs 531

and ice-fall-streaks are limited, more observations should be performed aimed at investigation of 532

the cloud processes (aggregation, riming, ice multiplication) in order to obtain a deeper 533

understanding of cloud microphysics. Except for observing real GCs, they could also be created 534

and observed in labs. Comprehemsive datasets, which could describe the structure and evolution 535

of such clouds could be used for the validation of complex microphysics models [Stoelinga et al., 536

2003; Keeler et al., 2016a]. 537

538

539

540

Acknowledgements 541

23

The APPRAISE-Clouds programme was funded by NERC (grant NE/E01125X/1). Radar and 542

aircraft datasets were made available by the British Atmospheric Data Centre (BADC). We would 543

like to acknowledge the support from FAAM and Direct Flight in obtaining the aircraft dataset. 544

ECMWF Interim Re-Analysis archive data were used for the brief presentation of the synoptic 545

condition of the present case (http://data.ecmwf.int/data/). Nimrod rain rate data were obtained 546

from UK CEDA (http://catalogue.ceda.ac.uk/). We would like to thank anonymous reviewers and 547

Prof David Schultz for the valuable feedback they provided, which improved the quality of the 548

paper. Finally, Mr Keppas was funded by the University of Manchester. 549

550

References 551

Andrić, Jelena, et al. "Polarimetric signatures above the melting layer in winter storms: An 552

observational and modeling study." Journal of Applied Meteorology and Climatology 52.3 553

(2013): 682-700. 554

Bader, M. J., S. A. Clough, and G. P. Cox. "Aircraft and dual polarization radar observations of 555

hydrometeors in light stratiform precipitation." Quarterly Journal of the Royal 556

Meteorological Society 113.476 (1987): 491-515. 557

Bailey and Hallett, 2009: A comprehensive habit diagram for atmospheric ice crystals: 558

confirmation from the laboratory, AIRS II, and other field studies, J. Atmos. Sci., 66 (9) 559

(2009), pp. 2888–2899 560

Balakrishnan, N., and D. S. Zrnic. "Estimation of rain and hail rates in mixed-phase precipitation." 561

J. Atmos. Sci 47.5 (1990): 561083. 562

Baumgardner, D., et al. "The cloud, aerosol and precipitation spectrometer: a new instrument for 563

cloud investigations." Atmospheric research 59 (2001): 251-264. 564

24

Bechini, R., L. Baldini, and V. Chandrasekar. "Polarimetric radar observations in the ice region of 565

precipitating clouds at C-band and X-band radar frequencies." Journal of Applied 566

Meteorology and Climatology 52.5 (2013): 1147-1169. 567

Boodoo, S., et al. "Application of dual-polarization radar melting-layer detection algorithm." 568

Journal of Applied Meteorology and Climatology 49.8 (2010): 1779-1793. 569

Brandes, Edward A., and Kyoko Ikeda. "Freezing-level estimation with polarimetric radar." 570

Journal of Applied Meteorology 43.11 (2004): 1541-1553. 571

Bringi, V. N., R. M. Rasmussen, and J. Vivekanandan. "Multiparameter radar measurements in 572

Colorado convective storms. Part I: Graupel melting studies." Journal of the atmospheric 573

sciences 43.22 (1986): 2547563. 574

Browning, K. A. "Structure of the atmosphere in the vicinity of large‐ amplitude Kelvin‐575

Helmholtz billows." Quarterly Journal of the Royal Meteorological Society 97.413 (1971): 576

283-299. 577

Browning, K. A., et al. "Air motion within Kelvin-Helmholtz billows determined from 578

simultaneous Doppler radar and aircraft measurements." QJR Meteorol. Soc 99 (1973): 579

608-618. 580

Browning, Keith A. "Conceptual models of precipitation systems." Weather and Forecasting 1.1 581

(1986): 23-41. 582

Browning, K. A. "The sting at the end of the tail: Damaging winds associated with extratropical 583

cyclones." Quarterly Journal of the Royal Meteorological Society 130.597 (2004): 37899. 584

Browning, K. A., and R. Wexler. "The determination of kinematic properties of a wind field using 585

Doppler radar." Journal of Applied Meteorology 7.1 (1968): 10613. 586

25

Chapman, D., and K. A. Browning. "Use of wind-shear displays for Doppler radar data." Bulletin 587

of the American Meteorological Society 79.12 (1998): 2687691. 588

Chen, Sue, and William R. Cotton. "The sensitivity of a simulated extratropical mesoscale 589

convective system to longwave radiation and ice-phase microphysics." Journal of the 590

atmospheric sciences 45.24 (1988): 3897-3910. 591

Choularton, T.W., Latham, J., Mason, B.J., 1978. A possible mechanism of ice splinter production 592

during riming. Nature 274, 791– 792. 593

Choularton, T.W., Griggs, D.J., Humood, B.Y., Latham, J., 1980. Laboratory studies of riming 594

and its relation to ice splinter production. Q. J. R. Meteorol. Soc. 106, 367– 373. 595

Colle, Brian A., Clifford F. Mass, and Bradley F. Smull. "An observational and numerical study 596

of a cold front interacting with the Olympic Mountains during COAST IOP5." Monthly 597

weather review 127.6 (1999): 1310-1334. 598

Conway, John W., and Dušan S. Zrnic. "A study of embryo production and hail growth using dual-599

Doppler and multiparameter radars." Monthly weather review 121.9 (1993): 2511-2528. 600

Crawford, I., et al. "Ice formation and development in aged, wintertime cumulus over the UK: 601

observations and modelling." Atmospheric Chemistry and Physics 12.11 (2012): 4963-602

4985. 603

Crosier, J., et al. "Observations of ice multiplication in a weakly convective cell embedded in 604

supercooled mid-level stratus." Atmospheric Chemistry and Physics 11.1 (2011): 257-273. 605

Crosier, J., et al. "Microphysical properties of cold frontal rainbands†." Quarterly Journal of the 606

Royal Meteorological Society 140.681 (2014): 1257-1268. 607

26

Dearden, C., et al. "Exploring the diabatic role of ice microphysical processes in two North 608

Atlantic summer cyclones." Monthly Weather Review 2016 (2016). 609

Dong, Ya Yi, and John Hallett. "Droplet accretion during rime growth and the formation of 610

secondary ice crystals." Quarterly Journal of the Royal Meteorological Society 115.485 611

(1989): 127-142. 612

Douglas, Richard Herbert, Kenrick Lewis Stuart Gunn, and J. Stewart Marshall. "Pattern in the 613

vertical of snow generation." Journal of Meteorology 14.2 (1957): 95-114. 614

Doviak, Richard J., Dusan S. Zrnic, and Dale S. Sirmans."Doppler weather radar." Proceedings of 615

the IEEE 67.11 (1979): 1522-1553. 616

Eckhardt, Sabine, et al. "A 15-year climatology of warm conveyor belts." Journal of climate 17.1 617

(2004): 218-237. 618

Ellis, Scott, et al. "Towards the detection of aircraft icing conditions using operational dual-619

polarimetric radar." 7 th European Radar Conference on Radar in Meteorology and 620

Hydrology, Toulouse, FR. 2012. 621

Fabry, Frederic, and Isztar Zawadzki. "Long-term radar observations of the melting layer of 622

precipitation and their interpretation." Journal of the atmospheric sciences 52.7 (1995): 623

838-851. 624

Field, P. R., A. J. Heymsfield, and A. Bansemer. "Shattering and particle interarrival times 625

measured by optical array probes in ice clouds." Journal of Atmospheric and Oceanic 626

Technology 23.10 (2006): 1357-1371. 627

Findeisen, W. "The formation of the 0 C isothermal layer and fractocumulus under nimbostratus." 628

Meteor. Z 57 (1940): 49-54. 629

27

Forbes, Richard M., and Peter A. Clark. "Sensitivity of extratropical cyclone mesoscale structure 630

to the parametrization of ice microphysical processes." Quarterly Journal of the Royal 631

Meteorological Society 129.589 (2003): 1123-1148. 632

Fukao, Shoichiro, Kyosuke Hamazu, and Richard J. Doviak. Radar for meteorological and 633

atmospheric observations. Springer Japan, 2014. 634

Giangrande, Scott E., and Alexander V. Ryzhkov. "Calibration of dual-polarization radar in the 635

presence of partial beam blockage." Journal of atmospheric and oceanic technology 22.8 636

(2005): 1156-1166. 637

Giangrande, Scott E., John M. Krause, and Alexander V. Ryzhkov. "Automatic designation of the 638

melting layer with a polarimetric prototype of the WSR-88D radar." Journal of Applied 639

Meteorology and Climatology 47.5 (2008): 1352364. 640

Godard, Serge. "Propagation of centimeter and millimeter wavelengths through precipitation." 641

IEEE Transactions on Antennas and Propagation 18.4 (1970): 530-534. 642

Goddard, J. W. F., J. D. Eastment, and J. Tan. "Self-consistent measurements of differential phase 643

and differential reflectivity in rain." Geoscience and Remote Sensing Symposium, 1994. 644

IGARSS'94. Surface and Atmospheric Remote Sensing: Technologies, Data Analysis and 645

Interpretation., International. Vol. 1. IEEE, 1994. 646

Griffin, Erica M., et al. "A polarimetric and microphysical investigation of the Northeast blizzard 647

of 8–9 February 2013." Weather and Forecasting 29.6 (2014): 1271-1294. 648

Hall, Martin PM, John WF Goddard, and Stephen M. Cherry. "Identification of hydrometeors and 649

other targets by dual-polarization radar." Radio Science (ISSN 0048-6604) 19 (1984): 132-650

140. 651

28

Hallett, J. and S. C. Mossop, 1974: Production of secondary ice particles during the riming process. 652

Nature, 249, 26–28. 653

Harrold, T. W. "Mechanisms influencing the distribution of precipitation within baroclinic 654

disturbances." Quarterly Journal of the Royal Meteorological Society 99.420 (1973): 232-655

251. 656

Herzegh, Paul H., and Peter V. Hobbs. "The mesoscale and microscale structure and organization 657

of clouds and precipitation in midlatitude cyclones. II: Warm-frontal clouds." Journal of 658

the Atmospheric Sciences 37.3 (1980): 597-611. 659

Herzegh, Paul H., and Peter V. Hobbs. "The mesoscale and microscale structure and organization 660

of clouds and precipitation in midlatitude cyclones. IV: Vertical air motions and 661

microphysical structures of prefrontal surge clouds and cold-frontal clouds." Journal of the 662

Atmospheric Sciences 38.8 (1981): 1771-1784. 663

Herzegh, Paul H., and Arthur R. Jameson. "Observing precipitation through dual-polarization 664

radar measurements." Bulletin of the American Meteorological Society 73.9 (1992): 665

1366374. 666

Heymsfield, Andrew. "Ice crystal terminal velocities." Journal of the Atmospheric Sciences 29.7 667

(1972): 1348-1357. 668

Heymsfield, Andrew J., and Jean Iaquinta. "Cirrus crystal terminal velocities." Journal of the 669

Atmospheric Sciences 57.7 (2000): 916-938. 670

Heymsfield, Andrew J., et al. "Refinements to ice particle mass dimensional and terminal velocity 671

relationships for ice clouds. Part II: Evaluation and parameterizations of ensemble ice 672

29

particle sedimentation velocities." Journal of the atmospheric sciences 64.4 (2007): 1068-673

1088. 674

Hobbs, Peter Victor, and Robert J. Farber. Fragmentation of ice particles in clouds. 1972. 675

Hobbs, Peter V., and John D. Locatelli. "Rainbands, precipitation cores and generating cells in a 676

cyclonic storm." Journal of the atmospheric Sciences 35.2 (1978): 230-241. 677

Hogan, Robin J., et al. "Properties of embedded convection in warm-frontal mixed-phase cloud 678

from aircraft and polarimetric radar." Quarterly Journal of the Royal Meteorological 679

Society 128.580 (2002): 451-476. 680

Houze Jr, Robert A., John D. Locatelli, and Peter V. Hobbs. "Dynamics and Cloud Microphysies 681

of the Rainbands in an Occluded Frontal System1." Journal of the Atmospheric Sciences 682

33.10 (1976): 1921-1936. 683

Houze Jr, Robert A. Cloud dynamics. Vol. 104. Academic press, 2014. 684

Hubbert, J., and V. N. Bringi. "An iterative filtering technique for the analysis of copolar 685

differential phase and dual-frequency radar measurements." Journal of Atmospheric and 686

Oceanic Technology 12.3 (1995): 643-648. 687

Hubbert, J., et al. "CSU-CHILL polarimetric radar measurements from a severe hail storm in 688

eastern Colorado." Journal of Applied Meteorology 37.8 (1998): 749-775. 689

Iguchi, Takamichi, et al. "WRF–SBM Simulations of Melting-Layer Structure in Mixed-Phase 690

Precipitation Events Observed during LPVEx." Journal of Applied Meteorology and 691

Climatology 53.12 (2014): 2710-2731. 692

30

Illingworth, A. J., J. W. F. Goddard, and S. M. Cherry, 1987: Polarization radar studies of 693

precipitation development in convective storms. Quart. J. Roy.Meteor.Soc., 113, 469–489. 694

Jayaweera, K. O. L. F., and R. E. Cottis. "Fall velocities of plate‐ like and columnar ice crystals." 695

Quarterly Journal of the Royal Meteorological Society 95.406 (1969): 703-709. 696

Jiang, Hongli, et al. "Cloud resolving simulations of mixed-phase Arctic stratus observed during 697

BASE: Sensitivity to concentration of ice crystals and large-scale heat and moisture 698

advection." Journal of the atmospheric sciences 57.13 (2000): 2105-2117. 699

Johnson, D. A., and J. Hallett. "Freezing and shattering of supercooled water drops." Quarterly 700

Journal of the Royal Meteorological Society 94.402 (1968): 468-482. 701

Kajikawa, Masahiro. "Measurement of falling velocity of individual snow crystals." Journal of the 702

Meteorological Society of Japan. Ser. II 50.6 (1972): 577-584. 703

Keeler, Jason M., et al. "Dynamics of cloud-top generating cells in winter cyclones. Part I: 704

Idealized simulations in the context of field observations." Journal of the Atmospheric 705

Sciences 73.4 (2016a): 1507-1527. 706

Keeler, Jason M., et al. "Dynamics of cloud-top generating cells in winter cyclones. Part II: 707

Radiative and instability forcing." Journal of the Atmospheric Sciences 73.4 (2016b): 708

1529-1553. 709

Keeler, Jason M., et al. "Dynamics of Cloud-Top Generating Cells in Winter Cyclones. Part III: 710

Shear and Convective Organization." Journal of the Atmospheric Sciences 74.9 (2017): 711

2879-2897. 712

Kemppi, Mirja L., and Victoria A. Sinclair. "Structure of a warm front: Helsinki testbed 713

observations and model simulation." Monthly Weather Review 139.9 (2011): 2876-2900. 714

31

Kennedy, Patrick C., and Steven A. Rutledge. "S-band dual-polarization radar observations of 715

winter storms." Journal of Applied Meteorology and Climatology 50.4 (2011): 844-858. 716

Keppas, S. Ch, et al. "Ice lollies: An ice particle generated in supercooled conveyor belts." 717

Geophysical Research Letters (2017). 718

Keyser, Daniel. "Atmospheric fronts: An observational perspective." Mesoscale Meteorology and 719

Forecasting. American Meteorological Society, 1986. 216-258. 720

Knight, Charles A. "Observations of the morphology of melting snow." Journal of the Atmospheric 721

Sciences 36.6 (1979): 1123-1130. 722

Knight, Charles A. "Ice Growth from the Vapor at-5°C." Journal of the Atmospheric Sciences 69.6 723

(2012): 2031-2040. 724

Knollenberg, Robert G. "The optical array: An alternative to scattering or extinction for airborne 725

particle size determination." Journal of Applied Meteorology 9.1 (1970): 86-103. 726

Kobayashi, T. "The growth of snow crystals at low supersaturations." Philosophical Magazine 727

6.71 (1961): 1363-1370. 728

Kollias, Pavlos, et al. "Millimeter-wavelength radars: New frontier in atmospheric cloud and 729

precipitation research." Bulletin of the American Meteorological Society 88.10 (2007): 730

1608-1624. 731

Koop, Thomas, et al. "Water activity as the determinant for homogeneous ice nucleation in 732

aqueous solutions." Nature 406.6796 (2000): 611-614. 733

Korolev, Alexei V., et al. "Microphysical characterization of mixed-phase clouds." Quart. J. 734

Roy.Meteor.Soc., 129.587 (2003): 39-65. 735

32

Kowalewski, Stefan, and Gerhard Peters. "Analysis of Z-R Relations Based on LDR Signatures 736

within the Melting Layer." Journal of Atmospheric and Oceanic Technology 27.9 (2010): 737

1556561. 738

Kumjian, MATTHEW R. "Principles and applications of dual-polarization weather radar. Part I: 739

Description of the polarimetric radar variables." J. Oper. Meteor 1.19 (2013): 226-242. 740

Kumjian, Matthew R., and Alexander V. Ryzhkov. "Polarimetric signatures in supercell 741

thunderstorms." Journal of Applied Meteorology and Climatology 47.7 (2008): 1940-1961. 742

Kumjian, Matthew R., et al. "High-resolution polarimetric radar observations of snow-generating 743

cells." Journal of Applied Meteorology and Climatology 53.6 (2014): 16315658. 744

Lance, S., et al. "Water droplet calibration of the Cloud Droplet Probe (CDP) and in-flight 745

performance in liquid, ice and mixed-phase clouds during ARCPAC." Atmospheric 746

Measurement Techniques 3.6 (2010): 1683-1706. 747

Langleben, Manuel Phillip. "The plan pattern of snow echoes at the generating level." Journal of 748

Meteorology 13.6 (1956): 554-560. 749

Lavers, David A., et al. "Winter floods in Britain are connected to atmospheric rivers." 750

Geophysical Research Letters 38.23 (2011). 751

Laws, J. Otis. "Measurements of the fall‐ velocity of water‐ drops and raindrops." Eos, 752

Transactions American Geophysical Union 22.3 (1941): 709-721. 753

Lawson, R. Paul, et al. "The 2D-S (stereo) probe: Design and preliminary tests of a new airborne, 754

high-speed, high-resolution particle imaging probe." Journal of Atmospheric and Oceanic 755

Technology 23.11 (2006): 1462-1477. 756

33

Lloyd, G., et al. "Observations of the origin and distribution of ice in cold, warm, and occluded 757

frontal systems during the DIAMET Campaign." Monthly Weather Review 142.11 (2014): 758

4230-4255. 759

Magono, C., and C. W. Lee, 1966: Meteorological classification of natural snow crystals. J. Fac. 760

Sci., Hokkaido Univ., Ser. 7, 2, 321–335 761

Marshall, John Stewart. "Precipitation trajectories and patterns." Journal of Meteorology 10.1 762

(1953): 25-29. 763

Matejka, Thomas J., Robert A. Houze, and Peter V. Hobbs. "Microphysics and dynamics of clouds 764

associated with mesoscale rainbands in extratropical cyclones." Quart. J. Roy.Meteor.Soc., 765

106.447 (1980): 29-56. 766

Matrosov, S. Y., 2008: Assessment of radar signal attenuation caused by the melting hydrometeor 767

layer. IEEE Trans. Geosci. Remote Sens., 46, 1039–1047, 768

doi:10.1109/TGRS.2008.915757. 769

Mitchell, David L. "Use of mass-and area-dimensional power laws for determining precipitation 770

particle terminal velocities." Journal of the atmospheric sciences 53.12 (1996): 1710-1723. 771

Moisseev, Dmitri N., et al. "Dual‐ polarization radar signatures in snowstorms: Role of snowflake 772

aggregation." Journal of Geophysical Research: Atmospheres 120.24 (2015): 126422655. 773

Morrison, Hugh, et al. "Possible roles of ice nucleation mode and ice nuclei depletion in the 774

extended lifetime of Arctic mixed‐ phase clouds." Geophysical research letters 32.18 775

(2005). 776

Mossop, S. C., 1976: Production of secondary ice particles during the growth of graupel by riming. 777

Quart. J. Roy. Meteor. Soc.,102, 45–57. 778

34

Mossop, S. C. "The influence of drop size distribution on the production of secondary ice particles 779

during graupel growth." Quart. J. Roy.Meteor.Soc., 104.440 (1978): 323-330. 780

Mossop, S. C. "Secondary ice particle production during rime growth- The effect of drop size 781

distribution and rimer velocity." Quart. J. Roy.Meteor.Soc., 111 (1985): 1113-1124. 782

Mossop, S. C. and J. Hallett, 1974: Ice crystal concentration in cumulus clouds: Influence of the 783

drop spectrum. Science, 186, 632–634. 784

Murakami, Masataka, et al. "Microphysical structures of warm-frontal clouds." Journal of the 785

Meteorological Society of Japan. Ser. II 70.5 (1992): 877-895. 786

Murray, B.J., O’Sullivan, D., Atkinson, J.D., Webb, M.E., 2012. Ice nucleation by particles 787

immersed in supercooled cloud droplets. Chem. Soc. Rev. 41, 6519–6554. 788

Oraltay, R. G., and J. Hallett. "The melting layer: A laboratory investigation of ice particle melt 789

and evaporation near 0 C." Journal of Applied Meteorology 44.2 (2005): 206-220. 790

Oue, Mariko, et al. "X-band polarimetric and Ka-band Doppler spectral radar observations of a 791

graupel-producing Arctic mixed-phase cloud." Journal of Applied Meteorology and 792

Climatology 54.6 (2015): 1336351. 793

Pfahl, Stephan, et al. "Warm conveyor belts in the ERA-Interim dataset (1979–2010). Part II: 794

Moisture origin and relevance for precipitation." Journal of Climate 27.1 (2014): 27-40. 795

Pinto, James O. "Autumnal mixed-phase cloudy boundary layers in the Arctic." Journal of the 796

atmospheric sciences 55.11 (1998): 2016-2038. 797

35

Pitertsev, Alexander, and Felix Yanovsky. "Polarimetric method for remote predicting a zone of 798

icing-in-flight in clouds and precipitation." Microwaves, Radar and Remote Sensing 799

Symposium (MRRS), 2011. IEEE, 2011. 800

Plummer, David M., et al. "Structure and statistical analysis of the microphysical properties of 801

generating cells in the comma head region of continental winter cyclones." Journal of the 802

Atmospheric Sciences 71.11 (2014): 4181-4203. 803

Plummer, David M., et al. "Microphysical properties of convectively generated fall streaks within 804

the stratiform comma head region of continental winter cyclones." Journal of the 805

Atmospheric Sciences 72.6 (2015): 2465-2483. 806

Protat, Alain, and Christopher R. Williams. "The accuracy of radar estimates of ice terminal fall 807

speed from vertically pointing Doppler radar measurements." Journal of Applied 808

Meteorology and Climatology 50.10 (2011): 2120-2138. 809

Pruppacher, H. R., and R. J. Schlamp. "A wind tunnel investigation on ice multiplication by 810

freezing of waterdrops falling at terminal velocity in air." Journal of Geophysical Research 811

80.3 (1975): 380-386. 812

Rauber, Robert M., et al. "The role of cloud-top generating cells and boundary layer circulations 813

in the finescale radar structure of a winter cyclone over the Great Lakes." Monthly Weather 814

Review 143.6 (2015): 2291-2318. 815

Rosenow, Andrew A., et al. "Vertical velocity and physical structure of generating cells and 816

convection in the comma head region of continental winter cyclones." Journal of the 817

Atmospheric Sciences 71.5 (2014): 1538-1558. 818

36

Ryzhkov, A. V., and D. S. Zrnic. "Discrimination between rain and snow with a polarimetric 819

radar." Journal of Applied Meteorology 37.10 (1998): 1228-1240. 820

Ryzhkov, A., S. Giangrande, and D. Zrnic. "Using multiparameter data to calibrate polarimetric 821

weather radars in the presence of a partial beam blockage." Geoscience and Remote 822

Sensing Symposium, 2002. IGARSS'02. 2002 IEEE International. Vol. 5. IEEE, 2002. 823

Ryzhkov, Alexander V., et al. "Polarimetric tornado detection." Journal of applied meteorology 824

44.5 (2005): 557-570. 825

Ryzhkov, A., et al. "Investigations of polarimetric radar signatures in winter storms and their 826

relation to aircraft icing and freezing rain." AMS 35 th Conf. on Radar Meteorology, 827

Pittsburgh, PA. 2011. 828

Sauvageot, H, Kouadio, K. and Etty, C. ‘The influence of temperatures and supersaturation on the 829

polarization of radar signals’. Pp. 173-176 in Volume 1, 23rd Conference on Radar 830

Meteorology, 22-26 Sept. 1986, Snowmass, Co. Am. Meteorol. SOC. 831

Seliga, T. A., and V. N. Bringi. "Potential use of radar differential reflectivity measurements at 832

orthogonal polarizations for measuring precipitation." Journal of Applied Meteorology 833

15.1 (1976): 69-76. 834

Serke, David J., et al. "The use of x-band radar to support the detection of in-flight icing hazards 835

by the NASA Icing Remote Sensing System." Remote Sensing Applications for Aviation 836

Weather Hazard Detection and Decision Support. Vol. 7088. International Society for 837

Optics and Photonics, 2008. 838

37

Schrom, Robert S., Matthew R. Kumjian, and Yinghui Lu. "Polarimetric radar signatures of 839

dendritic growth zones within Colorado winter storms." Journal of Applied Meteorology 840

and Climatology 54.12 (2015): 2367388. 841

Schrom, Robert S., and Matthew R. Kumjian. "Connecting microphysical processes in Colorado 842

winter storms with vertical profiles of radar observations." Journal of Applied Meteorology 843

and Climatology 2016 (2016). 844

Schuur, Terry J., et al. "Precipitation Observations with NSSL’s X-band Polarimetric Radar during 845

the SNOW-V10 Campaign." Pure and Applied Geophysics 171.1-2 (2014): 9612. 846

Smith, C. J., 1986: The reduction of errors caused by bright bands in quantitative rainfall 847

measurements made using radar. J. Atmos. Oceanic Technol., 3, 129–141, 848

doi:10.1175/1520-0426(1986)003,0129:TROECB.2.0.CO;2. 849

Smith, Paul L., et al. "Observations of mixed-phase precipitation within a CaPE thunderstorm." 850

Journal of Applied Meteorology 38.2 (1999): 14655. 851

Stark, David, Brian A. Colle, and Sandra E. Yuter. "Observed microphysical evolution for two 852

East Coast winter storms and the associated snow bands." Monthly Weather Review 141.6 853

(2013): 2037-2057. 854

Stewart, Ronald E. "Deep 0 C isothermal layers within precipitation bands over southern Ontario." 855

Journal of Geophysical Research: Atmospheres 89.D2 (1984): 2567-2572. 856

Stewart, Ronald E., et al. "Characteristics through the melting layer of stratiform clouds." Journal 857

of the atmospheric sciences 41.22 (1984): 3227-3237. 858

38

Stoelinga, Mark T., et al. "Improvement of microphysical parameterization through observational 859

verification experiment." Bulletin of the American Meteorological Society 84.12 (2003): 860

1807-1826. 861

Straka, Jerry M., Dusan S. Zrnic, and Alexander V. Ryzhkov. "Bulk hydrometeor classification 862

and quantification using polarimetric radar data: Synthesis of relations." Journal of Applied 863

Meteorology 39.8 (2000): 1341-1372. 864

Szeto, K. K., C. A. Lin, and R. E. Stewart, 1988: Mesoscale circulations forced by melting snow. 865

Part I: Basic simulations and dynamics. J. Atmos. Sci., 45, 1629–1641, doi:10.1175/ 1520-866

0469(1988)045,1629:MCFBMS.2.0.CO;2. 867

Szyrmer, Wanda, and Isztar Zawadzki. "Modeling of the melting layer. Part I: Dynamics and 868

microphysics." Journal of the atmospheric sciences 56.20 (1999): 3573-3592. 869

Giangrande, Scott E., John M. Krause, and Alexander V. Ryzhkov. "Automatic designation of the 870

melting layer with a polarimetric prototype of the WSR-88D radar." Journal of Applied 871

Meteorology and Climatology 47.5 (2008): 1354-1364. 872

Taylor, J. W., et al. "Observations of cloud microphysics and ice formation during COPE." 873

Atmospheric Chemistry and Physics 16.2 (2016): 799-826. 874

Vali, G., et al. "Technical Note: A proposal for ice nucleation terminology." Atmospheric 875

Chemistry and Physics 15.18 (2015): 10263-10270. 876

Vardiman, Larry. "The generation of secondary ice particles in clouds by crystal-crystal collision." 877

Journal of the Atmospheric Sciences 35.11 (1978): 2168-2180. 878

Wakimoto, Roger M., and Brian L. Bosart. "Airborne radar observations of a warm front during 879

FASTEX." Monthly weather review 129.2 (2001): 25374. 880

39

Werner, Frank D. "Gas temperature probe." U.S. Patent No. 2,970,475. 7 Feb. 1961. 881

Wexler, Raymond, and David Atlas. "Precipitation generating cells." Journal of Meteorology 16.3 882

(1959): 327-332. 883

Wolde, Mengistu, and Gabor Vali. "Polarimetric signatures from ice crystals observed at 95 GHz 884

in winter clouds. Part I: Dependence on crystal form." Journal of the atmospheric sciences 885

58.8 (2001): 828-841. 886

Zrnic´, D. S., N. Balakrishnan, C. L. Ziegler, V. N. Bringi, K. Aydin, and T. Matejka, 1993: 887

Polarimetric signatures in the stratiform region of a mesoscale convective system. J. 888

Appl.Meteor., 32, 678–693. 889

Zrnic, Dusan S., et al. "Testing a procedure for automatic classification of hydrometeor types." 890

Journal of Atmospheric and Oceanic Technology 18.6 (2001): 892-913. 891

892

40

Figure Captions 893

Figure 1. ECMWF Re-Analysis (ERA Interim) (0.125o resolution) for 21/01/2009 at 18UTC. (a) 894

Temperature (oC) on the 850mb pressure level (colour shading and dashed contours) and mean 895

sea level pressure (mb) (white contours). Surface fronts have been plotted in accordance with 896

UK MetOffice analysis charts. (b) Vertical section of relative humidity (colour shading) and 897

wind speed/direction (standard wind barbs) across (purple line in figure 1a) the investigated 898

warm front at 18z. Warm front position is highlighted by the red line according to wind and 899

potential temperature fields. 900

Figure 2. 3GHz radar RHI scans of (a) ZH (dBZ), (d) ZDR (dB) and (f) unfolded Doppler 901

velocity (m s-1) at 17:11:00-17:12:30UTC (negative values indicate flow towards the radar). (b) 902

Percentage of cloudy pixels (ZH dataset was used) with height. Graphs (c), (e), (g) and (h) show 903

average (solid lines) and 95th percentiles (dots) of ZH, ZDR, Doppler velocity and vertical wind 904

shear produced from Doppler velocity, which were calculated from the 3GHz radar dataset for 905

30≤Range≤90km and 0≤Altitude≤8km for the most representative scans during each time period 906

showed in the legend. Different colours indicate different time periods for (b), (c), (e), (g) and 907

(h). 908

Figure 3. (a) ZH (dBZ) and (b) Doppler velocity (m s-1) evolution with time as captured by the 909

35GHz radar. In (a) the melting layer height is highlighted by the blue (time <2100UTC), the 910

green (2100UTC< time <2215UTC) and the red (time <2215UTC) dashed lines. The black 911

dashed line shows the linear fitting of the cloud tops height between 1500-2100UTC. The 912

coloured lines on the left corner indicate the time that radiosondes of figure 4 were launched 913

comparing to the time colourscale of the figure 3a. In (b) negative Doppler velocity indicates 914

flow towards the radar. (c) Similar to figure 2b. (d) shows average (solid lines) and 95th 915

percentiles (dots) of ZH as calculated from 35GHZ radar data. Both (c) and (d) were calculated 916

from the first 10 minutes of data of each hour. 917

Figure 4. Graphs demonstrating (a) temperature (oC), (b) potential temperature (K), (c) relative 918

humidity (%), (d) wind speed (m s-1), (e) wind direction, (f) change of potential temperature 919

with height obtained from radiosonde data. Different colours indicate different times and were 920

selected to be comparable with the colour scheme of figures 2b, c, e, g and 3b, c, e. 921

Figure 5. Example of ice particle images captured by the 2D-S probe. Ice particles are grouped 922

by the temperature (left axis) and altitude (right axis) at which they were observed. 923

Figure 6. Unfolded Doppler velocity (m s-1) in RHI scans for three scanning periods: (a) 924

16:00:50 – 16:02:20, (c) 18:00:06 – 18:01:35, (e) 19:35:13 – 19:36:43 (the grayscale line shows 925

the aircraft track between 19:33:34 and 19:40:05 -black for the initial and white for the final 926

position). Negative values indicate flow towards the radar. The orange dashed line is the -20m s-927

1 contour. The solid red line in figure (c) determines the warm front boundary location and the 928

dashed red line the possible sub-front. (b) and (d) show radiosonde wind speed and direction 929

(colour scale) over the Chilbolton region (shaded curve) and Doppler velocity (blue dashed line) 930

along the vertical black line on graphs (a) and (c) correspondingly. (f) shows IWC, LWC and 931

cloud droplet concentrations for the corresponding aircraft track in (e) at T~ -1.9oC. 932

41

Figure 7. RHIs of ZH (dBZ): (a)-(d) and (i); and of ZDR (dB): (e)-(h) and (j). The horizontal 933

grayscale line shows the aircraft track between 19:03:00-19:07:48 in (i) and (j) (white for the 934

final, and black for the starting position). The mean temperature recorded along the track was ~ -935

7°C. The black contour line in (a)-(d) is for Doppler velocity equal to -20m s-1 and in (e)-(h) for 936

ZH=15dBZ. In figure (a) the red contour presents wind shear dV/dz=7m s-1 km-1. Figure (k) 937

shows LWC and cloud droplet concentration along the aircraft track in (i) and (j). 938

Figure 8. Probability distribution functions of different ZDR thresholds for in-situ (a) ice number 939

concentration, (b) mean ice particle diameter and (c) ZH. The data used for (a) and (b) come 940

from regions where the aircraft flew through ZDR/ZH fall-streaks. The data used for (c) come 941

from the 3GHz-radar for ranges of 30-90km and elevations >1km (to avoid the ML and focus on 942

fall-streaks) between 16:61:21-19:28:28 943

Figure 9. RHIs of (a)-(d) ZH (dBZ), (e)-(h) ZDR (dB). The grayscale line shows the aircraft 944

track between 16:21:13-16:24:22 in (c) and (g) and 16:52:39-16:59:06 in (d) and (h) (white for 945

the final, and black for the starting position). The black isoline in (a)-(d) is for Doppler velocity 946

equal to -20m s-1 and in (e)-(h) for ZH=15dBZ. Figures (i) and (j) show 2D-S images captured 947

during the flight track in (c), (g) and (d), (h) at temperatures -7.8°C and -12.9°C respectively. 948

Figures (k) and (l) show IWC/LWC and ice number concentration/ice mean diameter for the 949

aircraft track in (c), (g). Similarly, (l) and (n) correspond to the aircraft track in (d) and (h). 950

Figure 10. As in figure 9. Figures (i), (j), (k) and (l) show the surface rain rate (Nimrod) along 951

the radar range of the RHIs in (a,e), (b,f), (c,g) and (d,h) respectively. 952

Figure 11. (a) ZDR and (b) ZH profiles (0-1km height) from data taken from the radar scans 10d 953

and 10h. Green curves depict the profiles out (range=25km) and red curves within the ZH fall-954

streak (range=32km). 955

Figure 12. As in figure 9. In (g)) both 2D-S and CIP-100 images are shown. The time and range 956

from the Chilbolton radar are indicated at the bottom of the figure in order to be comparable with 957

the radar scans. In (h), the green curve shows the cloud droplet concentration. The mean 958

temperature along the aircraft track was ~ -4.8°C. 959

Figure 13. A schematic representation summarizing the investigated processes and polarimetric 960

signatures of the frontal clouds observed. 961

42

962

963

964

Figure 1. ECMWF Re-Analysis (ERA Interim) (0.125o resolution) for 21/01/2009 at 18UTC. (a) Temperature (oC) 965

on the 850mb pressure level (colour shading and dashed contours) and mean sea level pressure (mb) (white contours). 966

43

Surface fronts have been plotted in accordance with UK MetOffice analysis charts. (b) Vertical section of relative 967

humidity (colour shading) and wind speed/direction (standard wind barbs) across (purple line in figure 1a) the 968

investigated warm front at 18z. Warm front position is highlighted by the red line according to wind and potential 969

temperature fields. 970

971

972

Figure 2. 3GHz radar RHI scans of (a) ZH (dBZ), (d) ZDR (dB) and (f) unfolded Doppler velocity (m s-1) at 17:11:00-973

17:12:30UTC (negative values indicate flow towards the radar). (b) Percentage of cloudy pixels (ZH dataset was used) 974

with height. Graphs (c), (e), (g) and (h) show average (solid lines) and 95th percentiles (dots) of ZH, ZDR, Doppler 975

velocity and vertical wind shear produced from Doppler velocity, which were calculated from the 3GHz radar dataset 976

for 30≤Range≤90km and 0≤Altitude≤8km for the most representative scans during each time period showed in the 977

legend. Different colours indicate different time periods for (b), (c), (e), (g) and (h). 978

979

44

980

Figure 3. (a) ZH (dBZ) and (b) Doppler velocity (m s-1) evolution with time as captured by the 35GHz radar. In (a) 981

the melting layer height is highlighted by the blue (time <2100UTC), the green (2100UTC< time <2215UTC) and the 982

red (time <2215UTC) dashed lines. The black dashed line shows the linear fitting of the cloud tops height between 983

1500-2100UTC. The coloured lines on the left corner indicate the time that radiosondes of figure 4 were launched 984

comparing to the time colourscale of the figure 3a. In (b) negative Doppler velocity indicates flow towards the radar. 985

45

(c) Similar to figure 2b. (d) shows average (solid lines) and 95th percentiles (dots) of ZH as calculated from 35GHZ 986

radar data. Both (c) and (d) were calculated from the first 10 minutes of data of each hour. 987

988

Figure 4. Graphs demonstrating (a) temperature (oC), (b) potential temperature (K), (c) relative humidity (%), (d) 989

wind speed (m s-1), (e) wind direction, (f) change of potential temperature with height obtained from radiosonde data. 990

Different colours indicate different times and were selected to be comparable with the colour scheme of figures 2b, c, 991

e, g and 3b, c, e. 992

993

46

994

Figure 5. Example of ice particle images captured by the 2D-S probe. Ice particles are grouped by the temperature 995

(left axis) and altitude (right axis) at which they were observed. 996

997

47

998

Figure 6. Unfolded Doppler velocity (m s-1) in RHI scans for three scanning periods: (a) 16:00:50 – 16:02:20, (c) 999

18:00:06 – 18:01:35, (e) 19:35:13 – 19:36:43 (the grayscale line shows the aircraft track between 19:33:34 and 1000

19:40:05 -black for the initial and white for the final position). Negative values indicate flow towards the radar. The 1001

orange dashed line is the -20m s-1 contour. The solid red line in figure (c) determines the warm front boundary location 1002

and the dashed red line the possible sub-front. (b) and (d) show radiosonde wind speed and direction (colour scale) 1003

over the Chilbolton region (shaded curve) and Doppler velocity (blue dashed line) along the vertical black line on 1004

graphs (a) and (c) correspondingly. (f) shows IWC, LWC and cloud droplet concentrations for the corresponding 1005

aircraft track in (e) at T~ -1.9oC. 1006

1007

1008

1009

48

1010

Figure 7. RHIs of ZH (dBZ): (a)-(d) and (i); and of ZDR (dB): (e)-(h) and (j). The horizontal grayscale line shows the aircraft track between 19:03:00-19:07:48 in (i) and (j) (white 1011

for the final, and black for the starting position). The mean temperature recorded along the track was ~ -7°C. The black contour line in (a)-(d) is for Doppler velocity equal to -20m 1012

s-1 and in (e)-(h) for ZH=15dBZ. In figure (a) the red contour presents wind shear dV/dz=7m s-1 km-1. Figure (k) shows LWC and cloud droplet concentration along the aircraft track 1013

in (i) and (j).1014

49

1015

Figure 8. Probability distribution functions of different ZDR thresholds for in-situ (a) ice number concentration, (b) 1016

mean ice particle diameter and (c) ZH. The data used for (a) and (b) come from regions where the aircraft flew through 1017

ZDR/ZH fall-streaks. The data used for (c) come from the 3GHz-radar for ranges of 30-90km and elevations >1km (to 1018

avoid the ML and focus on fall-streaks) between 16:61:21-19:28:281019

50

1020

Figure 9. RHIs of (a)-(d) ZH (dBZ), (e)-(h) ZDR (dB). The grayscale line shows the aircraft track between 16:21:13-16:24:22 in (c) and (g) and 16:52:39-16:59:06 in (d) and (h) (white 1021

for the final, and black for the starting position). The black isoline in (a)-(d) is for Doppler velocity equal to -20m s-1 and in (e)-(h) for ZH=15dBZ. Figures (i) and (j) show 2D-S 1022

images captured during the flight track in (c), (g) and (d), (h) at temperatures -7.8°C and -12.9°C respectively. Figures (k) and (l) show IWC/LWC and ice number concentration/ice 1023

mean diameter for the aircraft track in (c), (g). Similarly, (l) and (n) correspond to the aircraft track in (d) and (h).1024

51

1025

Figure 10. As in figure 9. Figures (i), (j), (k) and (l) show the surface rain rate (Nimrod) along the radar range of the RHIs in (a,e), (b,f), (c,g) and (d,h) respectively. 1026

1027

52

1028

1029

Figure 11. (a) ZDR and (b) ZH profiles (0-1km height) from data taken from the radar scans 10d and 10h. Green curves 1030

depict the profiles out (range=25km) and red curves within the ZH fall-streak (range=32km). 1031

53

1032

Figure 12. As in figure 9. In (g)) both 2D-S and CIP-100 images are shown. The time and range from the Chilbolton radar are indicated at the bottom of the figure in order to be 1033

comparable with the radar scans. In (h), the green curve shows the cloud droplet concentration. The mean temperature along the aircraft track was ~ -4.8°C.1034

54

1035

Figure 13. A schematic representation summarizing the investigated processes and polarimetric signatures of the 1036

frontal clouds observed. 1037

1038