Numerical simulation of a rare winter hailstorm event over Delhi , India on 17 January 2013

This study analyzes the cause of the rare occurrence of a winter hailstorm over New Delhi/NCR (National Capital Region), India. The absence of increased surface temperature or low level of moisture incursion during winter cannot generate the deep convection required for sustaining a hailstorm. Consequently, NCR shows very few cases of hailstorms in the months of December-January-February, making the winter hail formation a question of interest. For this study, a recent winter hailstorm event on 17 January 2013 (16:00–18:00 UTC) occurring over NCR is investigated. The storm is simulated using the Weather Research and Forecasting (WRF) model with the Goddard Cumulus Ensemble (GCE) microphysics scheme with two different options: hail and graupel. The aim of the study is to understand and describe the cause of hailstorm event during over NCR with a comparative analysis of the two options of GCE microphysics. Upon evaluating the model simulations, it is observed that the hail option shows a more similar precipitation intensity with the Tropical Rainfall Measuring Mission (TRMM) observation than the graupel option does, and it is able to simulate hail precipitation. Using the modelsimulated output with the hail option; detailed investigation on understanding the dynamics of hailstorm is performed. The analysis based on a numerical simulation suggests that the deep instability in the atmospheric column led to the formation of hailstones as the cloud formation reached up to the glaciated zone promoting ice nucleation. In winters, such instability conditions rarely form due to low level available potential energy and moisture incursion along with upper level baroclinic instability due to the presence of a western disturbance (WD). Such rare positioning is found to be lowering the tropopause with increased temperature gradient, leading to winter hailstorm formation.


Introduction
An unusual winter hailstorm occurred over National Capital Region (NCR)/New Delhi, India on 17 January 2013 (16:00-18:00 UTC). Figure 1 shows satellite image of exten-25 sive cloud cover over the north Indian region during this time. Heavy precipitation was 6034 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | observed over north Indian region including some parts of western Himalayas. In NCR and surrounding regions along with severe rainfall there was freak hail incidence also. Table 1 corresponds to the summary of hailstorm cases over NCR (NNDC-CDO, 2013). Primarily, the Indian climate is divided into four seasons; pre-monsoon (March-April-May), monsoon (June-July-August-September), post-monsoon (October-November) 5 and winter (December-January-February) (Attri and Tyagi, 2010). With such a classification, a pattern is observed, where most hailstorms cases over NCR dominate during the warmer periods of pre-monsoon and monsoon months. It is seen that no hailstorm cases occurred during months of February and October whereas a very few hailstorm cases occur during the winter months. Out of the 33 hailstorm cases only 5 10 occur during the colder/winter period; with January showing only 2 cases. Thus, winter hailstorms can be considered a rare occurrence over NCR. With such a distinction between summer and winter hailstorms, it can be hypothesized that the mechanisms of these two kinds of hailstorms are different, which are deliberated in the following paragraphs. 15 Over north Indian region, pre-monsoon storm events occur during convectively unstable atmospheric conditions culminating due to transient disturbances observed in the air mass due to the surface heating. Storms may be categorized as severe storms if it is associated with hail, thunder, lightning, high winds etc (Houze Jr., 1981). These severe storms occur during strong vertical wind shear, which are ideal conditions for 20 hail formation (Orville and Kopp, 1977). During monsoon, the precipitation is caused due to development of deep convection because of higher surface temperature and vertical wind shear of south-westerly flow and associated moisture incursion over the Indian subcontinent. The convective precipitation is also facilitated by surface terrain and the Himalayan orography. These periods of precipitation occurrences during the 25 monsoon are associated with rainfall and in extreme cases hail (Koteswaram, 1958;Ramage, 1971;Sikka, 1977;Lau et al., 2012). In the context of hailstorms, it can be concluded that convection due to high surface temperatures and moisture laden flow are important for hailstorm formation. Hail is precipitation in the form of hard, rounded 6035 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | pellets of irregular lumps of ice. In cross section, hail shows concentric shells of different densities and degrees of opaqueness (Pruppacher and Klett, 2010). These multiple layers of ice within hail are formed due to continuous deposition and shedding of ice over the condensation nuclei as the hailstones cycle through strong convective clouds having multiple updrafts and downdrafts (Chatterjee et al., 2008). 5 The winter months are associated with absence of increased surface temperatures or low level of moisture incursion, making it a cold and dry season (Attri and Tyagi, 2010). These conditions are not conducive for generation of deep convection required for sustaining a hailstorm. In month of January the average temperature of Delhi is 13.5 • C (NNDC-CDO, 2013). With such low mean temperature, it is difficult for a storm 10 to attain the wind shear required for hail formation. Studying such unusual/abnormal meteorological condition conducive for winter hail formation is of interest. Thus, this study has a new and unique perspective to understand the dynamics of a winter hailstorm occurrence. Though hailstorms are a rare phenomenon due to very specific conditions of formation and subsequent occurrence (Table 1), they are quite hazardous. 15 Hailstones destroy crop, infrastructure, property and in extreme cases may cause injuries to humans (De et al., 2005). Due to the perspective of hailstorms happening only in summer, hailstorms are not expected in winter months. This may lead to unpreparedness in early warning systems for hail prediction and mitigation of damage due to hailstorms, which may lead to devastating consequences. 20 Numerical weather prediction technique is utilized for understanding the above discussed storm event. This storm is simulated with Weather Research and Forecasting (WRF) model with Advanced Research WRF (ARW) dynamical core. As hail formation is a microphysical process, the study mainly focuses on the use of Goddard Cumulus Ensemble (GCE) (Lin et al., 1983;Rutledge and Hobbs, 1984;Tao and Simpson, 25 1993) microphysics scheme to simulate hail formation. This scheme simulates six different hydrometeors or water particles: water vapour (WV), cloudwater (CW), cloudice (CI), rainwater (RW), snow and a third class of ice. The third class of ice can be graupel or hail as per the options used during simulation. Graupel is relatively smaller and less dense hydrometeor ice particle when compared with hail. These variations cause differences in microphysical properties of the model parameterization. This study includes analysis of simulations with both options of GCE microphysics for the hailstorm event, which will hereafter referred to as hail option or graupel option.
With these considerations, the objectives of this study are listed below:

5
-To understand and describe a winter hailstorm event over NCR, with discussion on the large scale flow and precipitation associated during such storms.
-To simulate the hailstorm event using GCE microphysics scheme's hail and graupel options for a comparative analysis in the two options.
-To study in detail the various thermodynamic and microphysical processes asso-10 ciated with the hailstorm formation using the GCE microphysics with hail option.
This paper is divided into the following sections with Sect. 2 describing experimental design and data, Sect. 3 showing results and discussion and summary and conclusion provided in Sect. 4.
2 Experimental design and data 15 In this study, the WRF model (version 3.0) with the ARW dynamic solver is used to simulate the hailstorm event. 2013, to understand the meteorological processes contributing to the storm development and propagation. In the current study, not only model performance in simulating the precipitation event is evaluated for verification, but the large scale atmospheric processes leading to the weather event are also analyzed. With such an aim, a three day simulation is deemed necessary for the study.

5
NCEP Final analysis data (FNL), at 1 • × 1 • spatial resolution, is used as the initial and lateral boundary conditions for the study (NCEP/NOAA/US-DoC, 2000). Initial conditions for the model are extracted from the FNL dataset and are interpolated to the model domain. NASA's Modern Era Retrospective-analysis for Research and Applications (MERRA) (Rienecker et al., 2011) 6 hourly analysis dataset with a spatial 10 resolution of 0.5 • × 0.7 • is used as corresponding observational data. Tropical Rainfall Measuring Mission (TRMM) multi-satellite precipitation analysis (real time data) 3B42 V7 (Huffman et al., 2007) at 0.25 • × 0.25 • spatial resolution is used as the observation for the precipitation fields. The daily OLR (Liebmann and Smith, 1996) dataset provided by Earth Systems Research Laboratory (ESRL), National Oceanic and Atmo- 15 spheric Administration (NOAA), USA, available at 2.5 • × 2.5 • spatial resolution is used for validation purposes.
As discussed previously the GCE microphysics scheme (Tao and Simpson, 1993) is used in the model simulation experiments. This parameterization scheme is based on the multi-dimensional, non-hydrostatic microphysics cloud resolving model (Simpson 20 and Tao, 1993). The microphysics scheme simulates six different hydrometeors: WV, CW, CI, RW, snow and the third class of ice (which can be either hail or graupel, as specified). The two options are different due to different microphysical processes considered for hail/graupel formation, which in turn impact the hydrometeor population. As both the hydrometeors are quite alike, their formation processes considered in both options of the parameterization scheme are similar; accretion of RW, aggregation of snow, rimming of CI, sublimation and melting (Lin et al., 1983). But with graupel option, two other processes are included for graupel formation; autoconversion of snow and deposition of WV (Rutledge and Hobbs, 1984). In this study to understand the microphysical process of hail formation and its complexities, simulation is conducted using the hail option specifically. It is to be mentioned that both simulations 3 and 1 km resolution nests were simulated with explicit representation of cumulus parameterization scheme. As model for simulations at horizontal resolutions smaller than 3 km, estimates the precipitation by the cloud microphysics scheme itself (Gomes and Chou,5 2010).

Analysis of large scale flow and precipitation
The model simulated large scale flow patterns at 500 hPa and its corresponding MERRA observation is depicted in Fig. 3. The large scale circulation showed a deep trough being formed over the western Indian and Pakistan region in model fields and corresponding MERRA observational analysis. Figure 3 depicts that both the model simulations capture the wind and geopotential height patterns with slight over estimation. The depression observed over the region corresponds to an incoming western disturbance (WD). WDs are eastward moving synoptic scale extra-tropical cyclonic 15 systems in the sub-tropical westerly jet. These originate in the Mediterranean Sea and cause precipitation over northern India mainly during the months of December-January-February, due to their interaction with the Himalayan orography (Dimri, 2004). This migratory disturbance in the mid-troposphere along with the stationary surface low over western India is called the WD, and generates the instability necessary for 20 winter precipitation (Dimri and Chevuturi, 2013). The presence of this system can be the cause for the enhanced the instability over the region that influences the storm formation.
Daily mean outgoing longwave radiation (OLR) for 17 January 2013 is depicted in Fig. 4

25
values is seen over whole of northern India including the NCR in both model options 6039 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | and observations. OLR can be considered as an indicator to describe the clouding condition or precipitation globally (Xie and Arkin, 1998). In the presence of increased extent and depth of cloud cover and lower cloud top temperatures in high clouds, low OLR values are recorded. Thus, the figure represents the cloud extent during the time period of the hailstorm. On evaluating the satellite image of Fig. 1, similar cloud extent is ob-5 served over the same region. The low values of OLR over the whole of northern Indian region depict extensive cloud cover and the possibility of precipitation over the region. The advance of WDs and interaction with Himalayas is associated with the development of mesoscale cloud formation (Puranik and Karekar, 2009) as seen during this event. Moisture flux transport and divergence integrated over the vertical atmospheric 10 column for observation and model simulations is shown in Fig. 5. In model simulated outputs, a strong region of moisture convergence is seen over NCR and some parts of Himalayan region north of NCR. Corresponding observations show a similar extent of moisture convergence zone, but the strength of moisture convergence is lower. According to the figure, it is observed that Arabian Sea and Bay of Bengal are both sources of 15 moisture for the convergence zone. The moisture laden flow from these sources flows over the Indian subcontinent, congregates and flows towards NCR. This moisture is important for the maintenance of the winter storm over northern India and its role will be discussed later in detail.
Analysis of three hourly precipitation rate from model simulations and its correspond- 20 ing observation is shown in Fig. 6. The figure suggests that there was a storm having maximum intensity over NCR between 15:00 to 18:00 UTC of 17 January 2013. This storm showed a localized formation over NCR and its surrounding region. When compared with the observation, model simulated output shows a similar spatial extent of precipitation patterns. The axis of the storm evolution is observed to be oriented along 25 north-east to south-west direction. On evaluating the two model simulations, it is observed that hail option shows closer precipitation intensity with TRMM observation than the graupel option. The spatial extent of precipitation pattern observed in hail option is also better match when compared with the observational analysis. Figure also depicts that storm cluster moves eastward as the storm progresses. The analysis shows localized cluster of storm that is the cause of the heavy precipitation event. But when the point specific precipitation intensity at NCR is compared between the different options, overestimation is observed in model simulation. Inter-comparison in temporal variation in half hourly precipitation intensities of both model simulations is shown in Fig. 7 for 5 all four horizontal model resolutions. The figure suggests peak concentration of precipitation around NCR region was observed around 17:00 UTC on 17 January 2013. Here it is also clearly noted that hail option shows higher precipitation estimates than the graupel option. Deliberations on reasons for variation in precipitation intensities within the options of microphysics are provided in next section by comparative analyses of 10 the flux in the different hydrometeors in different options.

Comparative analysis between hail and graupel options of GCE microphysics
Figure 8 depicts the mixing ratios of different hail and graupel outputs over spatial and temporal scales for both hail and graupel options. Increased concentrations of mixing 15 ratios of all six hydrometeors are observed over the NCR region around the peak of the storm (Supplement; Figs. S1-S5). These hydrometeors are indicators of cloud formation over the region. All the hydrometeor mixing ratio concentrations show higher values over NCR around 16:30 UTC, which is just before storm peaks representing storm evolution and/or build-up. In hail option, the increased formations of hydrometeors are 20 specifically observed over NCR, whereas some displacement is seen in graupel option output. In addition, the spatio-temporal variation of these mixing ratios shows an eastward movement of the storm. The increase of WV in the atmospheric column is due to the moisture incursion from the Arabian Sea and Bay of Bengal. As the WV rises in the atmosphere, formation of CW and CI begins by condensation and deposition pro-25 cesses due to decreasing temperatures, which form the clouds. The CW droplets are suspended in the atmospheric column in the mid-tropospheric layers, whereas CI particles are found in upper-troposphere where much colder conditions prevail. The RW 6041 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | droplets formation is enhanced by the processes of collision and coalescence (Oriville and Kopp, 1977), and ultimately sediment out as rain. Increased RW mixing ratios are an indicator of increased precipitable water availability in the atmospheric column. With the development of stronger storm conditions, there is formation of other hydrometeors like snow, hail and graupel. With the main focus of the paper on the third class of ice, 5 we will be discussing hail and graupel in more detail. Hail and graupel are both ice particles seen during storm formation. Hail is an ice particle which is formed due to consecutive processes like soft hail processes (dry growth and raindrop freezing) or hard hail processes (wet growth and shedding) (Wisner et al., 1972). These continuous and repeated processes form a differentially layered ice pre-10 cipitation type called hail (Pruppacher and Klett, 2010). The hailstones grow as the ice particles cycle through multiple cells of convective clouds (Chatterjee et al., 2008). Graupel are similar but smaller ice particles with a diameter less than 5 mm, formed exclusively by soft hail processes (Pruppacher and Klett, 2010). In the model simulation peak hail mixing ratios are observed over NCR around 16:00-16:30 UTC, whereas 15 graupel mixing ratios maxima is seen east of NCR. In the vertical profile hail mixing ratios span from 800-200 hPa whereas graupel mixing ratios show extent from 750-300 hPa. It is interesting to note that when hail and graupel mixing ratios are compared, the hail mixing ratios are lower than graupel mixing ratios. But graupel precipitation or sedimentation is not observed in the model simulation but hail precipitation is observed 20 in the hail option around NCR between 16:30-17:00 UTC as shown in Fig. 9. This observation can be attributed to the fact that graupel have high number concentrations in the atmospheric column as these small ice particles are formed much quicker than hailstones. But due to their small size the graupel particles also melt quicker due to the temperature conditions that are not as low as seen over the snowline. Thus, sedimentation of graupel is not observed over NCR but hail precipitation is seen during this hailstorm. With this discussion it can be concluded that to understand hail formation over NCR, microphysics with hail option needs to be studied in depth. Subsequent section focuses on understanding the cause of winter hailstorm formation.
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Detailed analysis of winter hailstorm formation
For a focused analysis of the winter hailstorm formation the time period of peak precipitation 17:00 UTC 17 January 2013 is considered (Fig. 10). Vertical cross sections of various parameters analyzed along the axis of core precipitation zone of the storm region as demarcated by a green line Fig. 10. Some of the model simulated fields are also 5 evaluated by considering their area averaged values over the 1 • × 1 • grey box around NCR as drawn in Fig. 10. The region over and around 77.2 • E and 28.6 • N would be considered the NCR or the region of study/interest. The analysis of the dynamical properties pertaining to the winter hailstorm along the axis demarcated in Fig. 10 is provided in Fig. 11. Model simulates a region of high 10 equivalent potential temperature (EPT) over the NCR, as seen in the Fig. 11a. This represents higher temperatures and moisture content in the mid-troposphere whereas in lower troposphere decreased EPT is noticed. This represents the region of instability spanning the atmospheric column which is the cause of the storm. The instability in the mid-tropospheric level is caused by the WD depression as observed at 500 hPa. 15 In this region, vertical wind (Fig. 11a) shows cells with updraft and downdraft conditions. Hailstorms forming processes are known for multiple cellular structure showing regions of updrafts and downdrafts within the cloud, with dynamic movement and reflectivity patterns. The updrafts originate from the air inflow from the surface towards the cloud in the opposite direction of storm movement (Chalon et al., 1976). Whereas, 20 the downdraft cells are caused by the air that ascends in the updrafts, and ultimately flows as outflow of air towards the front and back of the storm movement (Frankhauser, 1976;Strauch and Merrem, 1976). Relative vorticity and divergence in Fig. 11b, show alternative cells of positive and negative relative vorticity over NCR, which correspond to cyclonic and anti-cyclonic circulation respectively. These cells represent a gradient 25 of vorticity, depicting a positive vorticity advection which culminates in rising air. Associated to these cells are regions of divergence and convergence. The convergence zone observed in the mid-tropospheric level corresponds to the multiple cells of the 6043 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | hailstorm. These define the movement of hydrometeors through the associated updrafts and downdrafts. As discussed above, the multiple cells formed in the clouds promote the formation of hailstones. Further, positive specific humidity anomaly over NCR is observed in the mid-troposphere (Fig. 11c). This indicates that the atmospheric column over NCR upto 200 hPa shows a high moisture zone. This increased level of 5 moisture is required for not only growth of rain drops but also for diffusional and accretional growth of the hailstones. With the evolution of the storm, there is increase in different hydrometeors as discussed above. This indicates increase in reflectivity during storm formation (Fig. 11c). The storm shows increase of RW droplets, which congregate in the lower levels of atmosphere, showing high reflectivity values in the same To understand the reason for development of instability driving the winter hailstorm, 15 the geopotential height over the region is studied (Fig. 11d). Over NCR the geopotential height anomaly shows an increase around 400-200 hPa. This increase is associated with the dipping in the perturbation geopotential height contour lines. These changes are due to the tropopause fold penetrating the troposphere. The dip in the tropopause height values is also observed in the station data over New Delhi (Source: university 20 of Wyoming, Sounding data). This tropopause lowering is associated with baroclinic instability occurring over the region (Bush and Peltier, 1994). The increase storm intensity over the region is caused by the baroclinic instability due to the passing WD and the development of cyclonic circulation. The mid-latitude migratory WD attains higher intensities in form of a baroclinically unstable disturbance specifically over In-25 dian region (Rao and Rao, 1971;Singh and Agnihotri, 1977). This instability in the mid-to upper tropospheric levels generates the turbulent convective energy required for the development of updrafts during storm occurrence. With the availability of moisture in the atmospheric column the instability leads to heavy precipitation. But a WD over northern India does not always lead to hail formation during winter. Hail precipitation in model simulation is seen from 16:00-17:00 UTC 17 January 2013 (Fig. 9). The updrafts driven by the instability developed over the region, cycles the hail through the cloud. Hail particle successively moving through the vertical column grows in the upward motion and melts/sheds in the downward movement. With each cycle the hail-5 stone grows a new layer of ice forming the concentric circles seen in a hailstone cross section as discussed above. Thus, vertical wind velocity is an important factor for the hail formation. Heymsfield et al. (2005) describes that strong convective updrafts (with vertical wind speed greater than 5-10 m s −1 ) suppress homogenous nucleation to form ice particles which grow to form hail. Whereas, lower wind speeds would not attain 10 enough energy to develop a strong hailstorm. The model simulated vertical wind updraft speeds over NCR show a magnitude of 4-6 m s −1 which provide sufficient time for ice particle growth by dry or wet growth. The instability developed in the mid-tropospheric levels due to the WD develops propensity for baroclinic atmosphere in the upper half of troposphere. When the tem-15 poral variation of temperature profile of the region is analyzed, a dip is observed in the −60 • C isotherm around 16:00-17:00 UTC (Fig. 12). This lowering corresponds to the troposphere fold discussed before. The lowering of tropopause causes incursion of colder stratospheric layers into warmer troposphere. This in turn causes development of a steep temperature gradient as seen in the figure, which enhances upper level in-20 stability. Still the reasons for instability in the lower layers are yet not clearly discussed.
To understand the lower layer instability, temporal variation in area averaged convective available potential energy (CAPE) and specific humidity are represented in black and blue contours of Fig. 12 respectively. In this figure, an increase of moisture over NCR in the lower levels of atmospheric column is observed along with development of CAPE 25 from around 13:00 UTC. The source of this low level moisture incursion is majorly from Arabian Sea and on a lesser extent from Bay of Bengal (Fig. 5). The moisture convergence develops buoyancy which enhances the propensity of increase of CAPE in the atmospheric column. There is reduction in CAPE values in subsequent time periods 6045 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | after 13:00 UTC. The increase of CAPE defines the potential energy that is available to drive a storm and release of CAPE in form of kinetic energy promotes storm development. Along with this the low level moisture incursion provides the buoyancy required for the air parcel to rise. Thus, upper level instability is predominantly by the existing WD embedded with the low level instability due to moisture incursion and development 5 of CAPE lead to instability spanning the troposphere which makes it conducive conditions for hailstorm formation. Muller et al. (2008) and Rosenfeld and Lensky (1998) described the "continental" clouds having three zones in the vertical direction based on temperature variation: diffusional droplet growth zone (upto −10 • C), mixed phase zone (−10 to −38 • C) and glaciated zone (above −38 • C). The isotherm of −38 • C is termed 10 as homogeneous freezing isotherm beyond which homogenous nucleation occurs. As per Fig. 12, it can be concluded that the mixed phase zone for the current hailstorm can be considered within 600-350 hPa. The mixed phase zone is the region for development of hail particles due to growth by deposition of water particles. This zone in the model simulated output coincides with region having strong convergence influence 15 and higher specific humidity, promoting hail formation along with raindrop growth. The region of glaciation in the upper levels of troposphere is imperative for the development of small ice particles through homogenous nucleation. These particles further grow to form the various ice precipitation forms, in this case hail. The problem with winter storms is that instability is not sufficient enough for the cloud to extend to this height or 20 form an anvil. But in the 17 January 2013 hailstorm, the instability extending from low to upper levels of troposphere discussed above allows the formation of ice nuclei in the glaciations zone leading to winter hail formation. The clarity of two different zones of instability can be described in the sounding data of NCR for 17 January 2013. Figure 13 represents the model and observed sounding data in a graphical format. The development of instability at 12:00 UTC is an indicator for the storm development. This instability is measured by the gap between the air parcel lapse rate and environmental lapse rate. There is increased instability around mid-tropospheric level in both model and observation due to the approaching WD and development of baroclinic instability. Another region of higher instability seen in the observation is around 700 hPa, which is underestimated in the model simulation. This region of instability corresponding to the lower level convection described due to development of CAPE and the moisture incursion. There is augmented moisture content in the atmospheric column as seen in the relative humidity profile provides buoyancy 5 to the air parcel. Combination of low level available potential energy and moisture incursion along with upper level baroclinic instability due to the presence of WD, led to sufficient instability for winter hail formation.

Summary and conclusions
This study investigates the cause of hailstorm during winter over NCR. Winter months of December-January-February, specifically over north India, are considered a cold and dry season. But high surface temperature and supplementary moisture availability are the prerequisite for hailstorm development. And winter storms are not able to attain the intensity and depth of a convective storm. Thus, the question of hail being formed during the winter season is fascinating due to its uniqueness. 15 The comprehensive analysis of the results has led the authors to describe the mechanism of the winter hailstorm formation over NCR. These conclusions have been summarized in the form of an illustration or conceptual model as given in Fig. 14. On the day of the hailstorm, 17 January 2013, high moisture availability is seen over NCR. Arabian Sea and Bay of Bengal are identified as the source of this moisture in-flow. 20 This moisture incursion along with the release of CAPE in the lower levels leads to development of the lower level instability. But this instability is not sufficient for a winter storm. On the other hand winter precipitation is experienced over north India due to WDs. These migratory disturbances in the sub-tropical westerly jet cause baroclinic instability in the mid/upper troposphere. Thus, with two different sources of instability 25 this winter storm is able attain conditions similar to deep convective storms favorable for hail formation. The lower level instability formed due to release of kinetic energy and buoyancy of air parcel due to high relative humidity. This mid/upper level instability develops due to lowering of the tropopause which intensifies the temperature gradient. Thus, a rising air parcel gets an enhanced push to continue rising, which denotes the formation clouds having larger vertical extents. Ice nucleation, which develops hail nuclei, is enhanced 5 in such clouds as they attain the glaciated zone. These clouds are also associated with cells having turbulent upward and downward wind movement. Development of these conditions supports cycling of hail through the clouds and promotes its formation through shedding, deposition and other hail processes. As the hailstones grow, they ultimately sediment or precipitate out to cause the hailstorm during winter.

6065
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Figure 13. Sounding data at NCR at 12:00 UTC 17 January 2013 in graphical output for (a) observation and (b) 1 km horizontal model resolution with hail option.